Skip to main content

Full text of "The witness of the Vulgate, Peshitta and Septuagint to the text of Zephaniah"

See other formats







No. IV 


















No. IV 















Copyright, 1909 

Printed from type. Published May, 1909 



No complete examination of the relation of the chief Versions 
of the Old Testament to the original Hebrew has been made with 
especial reference to the Book of Zephaniah. Dr. Zandstra has 
in the following Essay supplied this want with much care and 

May 20th, 1909. 





I. The Vulgate, 

II. The Peshitta, 

III. The Septuagint, - 

IV. The Interdependence of the. Versions, 

- 1-6 

- 6-17 

- 18-24 

- 24-35 

- 35-38 

V. Departures from Massoretic Tradition and Variants 

from Consonantal Text, - 38-45 

VI. Conclusion, - - 45-47 

Appendix I. Difficulties in the Hebrew Text, - 4749 

Appendix II. Conjectural and Higher Criticism of the 

Text, - 49-52 


I. It is proposed in the following pages to study the text of 
Zephaniah in the light of the ancient primary versions. This 
study was undertaken largely to become familiar with Old Testa- 
ment Criticism a field of which it is peculiarly true that orien- 
tation is possible only at first hand. The choice of so short a 
text is vindicated by the almost unanimous verdict of scholars 
that the work of the translators of these versions is very uneven 
in quality. It is in fact still a moot question whether the Minor 
Prophets were translated into Greek by one individual or by 
many ; and the arguments that have been advanced 1 to show that 
the Peshitta is not really a deliberate translation, but rather the 
final stereotyped form that traditional renderings of various 
origins assumed, have never been satisfactorily met. The reasons 
for the choice of this particular text are two. (a.) Though the 
Hebrew of Zephaniah presents many difficulties, no complete 
study of its text corresponding to such work as has been done on 
Micah by Ryssel 2 seems ever to have been made, (b.) In critical 
commentaries it always occupies a subordinate place among the 
Minor Prophets, and in textual studies it is entirely overshadowed 
by the more important books of the division of the Canon to 
which it belongs. 3 This neglect, whatever its explanation may 
be, makes Zephaniah a good choice for a textual study. As it 
would be fatal presumption for one to ignore the work of prede- 
cessors, whether it bore directly or indirectly on one's theme, it 

1 Perles, Meletemata Peschittoniana, 1859, p. 48. 

a Ryssel, Untersuchungen uber die Textgestalt und die Echtheit des Buches Micha, 

aSchwally's Das Such Zephanja, Z.A.T.W. (1885), pp. 183 ff., is the only separate 
commentary outside of the well-known English and German critical series accessible to 
the general student. Bachmann has written specifically about the text of Zephaniah in 
an article entitled Zur Textkritik des Propheten Zephanja, S.K. (1894) ; his article is, 
however, but a statement of conclusions, and it is characterized by a most reckless spirit 
of conjecture. Here and there a brief note on some proposed emendation is to be 
found ; cf. Z.A.T.W. (1885), pp. 183 ff. and Z.A.T.W. (1891), pp. 185 f., 260 ff. 

2 The Text of Zephaniah. 

goes almost without saying that all available sources of informa- 
tion have been carefully examined and freely laid under tribute. 
That which is presented, while based on original investigation, 
has thus also of necessity the virtue of being a more or less com 
plete digest of the work of others. 1 

II. Because Old Testament Criticism is still for many reasons 
a wilderness through which each one must in large part blaze his 
own trail, it seems necessary to preface the statement of the 
method chosen in this examination by some more general remarks 
that shall not only explain it, but also justify its use. 

(A.) The thesis that all extant Hebrew sources for the text of 
the Old Testament, both in manuscript and in print, go back to 
a first century archetype, was first advanced by Lagarde in 1863. 
The chief supports of this thesis are the remarkable uniformity 
that is found in the manuscripts on the one hand, and the sup- 
posedly large number of corruptions in the text on the other. 
These two phenomena are mutually exclusive in an ancient docu- 
ment that has been accurately transmitted from its autograph, 
and their conjunction in this case is said to demand a comparatively 
late date for the common source to which all manuscripts and 
printed editions converge. The date of this hypothetical archetype 
is fixed in the first century by certain external characteristics 
that the text presents and by known facts in Jewish History. 3 
Strack, who about thirty years ago could pass over this view in 
silence, 3 states in his article on the Text of the Old Testament in 

1 A bibliography has not been prepared because complete lists of the literature that 
must be consulted abound. Berger (Histoire de la Vulgate pendant les premiers 
siecles du moyen dge), Swete (The Old Testament in Greek) and Nestle (Urtext und 
Ubersetzungen der Bibel, reprinted in the Real-Encyclop'adie fur protest. Theologie 
und Kirche) are practically exhaustive as far as the general literature is concerned. 
To the commentaries mentioned in Hastings' Dictionary of the Bible (article Zephaniah) 
those of Marti and Driver must be added ; in the miscellaneous literature Ehrlich 
(Mikrd Ki-Pheschut6, III, pp. 456-463) may well be included. This last work is written 
in Hebrew, but a German translation of the passages discussed is given. 

2 In a few characteristic paragraphs (Symmicta, II, pp. 120, 121), intended primarily 
to show that this thesis was entirely original with himself, Lagarde incidentally gives a 
brief account of how it had been received by scholars up to 1880. It appears that Ols- 
hausen had independently reached a very similar view through a different process of 
reasoning. Cf, further Swete, Introduction to the Old Testament in Greek, pp. 313-320 ; 
W. R. Smith, Old Testament in the Jewish Church, p. 56 ; Driver, Notes on the Hebrew 
Text of Samuel, pp. xxxix ff. 

8 Lagarde, Symmicta, II, p. 120. 

Introduction. 3 

Hastings 1 Dictionary of the jBible that it is accepted by most 
moderns. He himself does not accept it, but holds that the cus- 
tom of consigning manuscripts that had been damaged by the 
tooth of time, by fire, or by water, or that were found to contain 
more than a certain number of mistakes, to the so-called genizah, 
which was generally a room in the cellar of a synagogue, is suffi- 
cient to explain all the 1 phenomena. This thesis, whether true or 
not, offers striking proof that the present Hebrew text gives but 
scant aid in tracing its own history beyond a certain point, or in 
fixing its earliest form. Moreover, there are but few manuscripts, 
of which none are very old, and textual types the chief material 
for the criticism of texts are thus not to be found. 1 But it is a 
cardinal principle of criticism that to recover the true text of an 
ancient document it is first necessary to know its history ; and 
that manuscripts, although the text which they contain is undated 
and unlocalized, generally furnish the primary data for reconstruct- 
ing this history with the help of versions, which serve in a sec- 
ondary capacity to fix the time and place of origin of the differ- 
ent textual types that the manuscripts present. In the Old Tes- 
tament, however, there are no types of text in regard to which 
versions can be made to indicate a choice, but they themselves 
become the principal data. Instead of being called on to show 
from which particular type of two or more existing types it was 
made, a version must surrender the text on which it was based, 
in order that it may then be decided whether that text agrees 
with or differs from the single Hebrew textual type. Because a 
version must thus itself yield the text from which it was made, 
Old Testament Criticism is complicated by all the variable factors 
necessarily connected with translation and translators. 

(B.) Languages are for the most part so different in genius that 
translation from one into another is often impossible without theft 

1 Ginsburg's new ' Edition of the Hebrew Bible according to the Massoretic Text of 
Jacob Ben Chayim ' (British and Foreign Bible Society, August, 1908) contains the results 
of a collation of 71 manuscripts and 19 early printed editions. The editor has presuma- 
bly used everything that seemed worth using in this latest edition and yet there are at 
most but 27 manuscripts and 9 early printed editions of the Prophets cited. The earliest 
of the manuscripts is dated 916 A. D. Although sixth century dates have been defended 
for certain manuscripts, that of the Pentateuch from circa 820-850 (Or. 4445) and the 
Karaite synagogue manuscript of the Latter Prophets, 'written 827 years after the destruc- 
tion of the Temple,' i. e., 895 A. D., are generally regarded as the oldest. 

4 The Text of Zephaniah. 

from the thought of the first or assault upon the idiom of the 
second. The vagaries of translators are also all but incalculable. 
In testing one's retranslation of a reading the dividing line 
between the necessary use of the Hebrew text for guidance and 
prejudicial dependence upon it is hard to locate. Because he 
cannot entirely penetrate the structural difference of the two 
dead languages, the critic is inclined to find variants where none 
exist ; and in obvious disagreements he is apt to make too little 
allowance for the translator whose mental processes he cannot 
sufficiently follow, and whose knowledge and ability he cannot 
accurately gauge. Enough has been said to show that the 
"peculiarities of each translator, the character of his translation, 
and the knowledge of both languages displayed " by him infor- 
mation in regard to these matters can of course be gained only by 
comparisons both within and beyond the limits of the book being 
studied 1 are determining factors in the evaluation of his version. 
It is also evident that the large factor of ignorance by which the 
critic is necessarily handicapped establishes in all doubtful cases a 
strong presumption in favor of the agreement of the current 
Hebrew with the source of a version. 2 

(C.) The necessity of freeing the text of each version from 
inner corruptions by tracing it as far back as possible is patent. 
Neither the Vulgate, Peshitta nor Septuagint can, however, be 
carried back to the time of their origin, 3 and it is therefore neces- 
sary to seek such help as early quotations can give. The mutual 
relation of the versions has an important bearing on their value 
as witnesses, and consequently the presence or absence of inter- 
dependence must be established. 

1 In the case of the Septuagint these comparisons are much facilitated by the excel- 
lent concordances available, but with the Peshitta the work is most difficult because of 
the lack of these helps. Dutripon's Concordantiae Bibliorum Sacrorum Vulgatae 
Editionis can be used with great advantage together with a Hebrew concordance. 

2 Of the three equations Version <: Massoretic Text, Version = Massoretic Text and 
Version > Massoretic Text, the possibilities of the second must be exhausted before the 
others can present themselves. Ryssel assumed that the Massoretic Text was preferable 
to the Septuagint ; Frankel tried always to make the Massoretic Text equal the Septua- 
gint ; Streane held that the Septuagint was better than the Massoretic Text (cf. Stek- 
hoven, De Alexaandrijnsche Vertaling van het Dodekaprofeton, p. 121; Hastings' 
Dictionary of the Bible, IV, p. 731 b ). Frankel's results are therefore in so far forth the 
most dependable. 

8 It is not definitely known when the Septuagint and the Peshitta originated ; and 
although Jerome translated Zephaniah about 393 A. D., the date of the manuscripts used 
by him is unknown. 

Introduction. 5 

III. The method of procedure adopted in the present inquiry 
is based on the above considerations. The history of the versions 
has been separately discussed to locate and establish the best 
obtainable text of Zephaniah in each. The equivalents, which are 
obviously due to the character of the translation or to linguistic 
necessity, and those which must, because of the absence of evi- 
dence to the contrary, be ascribed to the characteristics or nuances 
of the translator, have been grouped together, and for the Vul- 
gate presented in a summary, for the Peshitta and Septuagint 
exhibited in toto. The question of interdependence has been 
considered, and such readings as have demanded individual con- 
sideration have been discussed. Thus the versions have been 
summoned to show cause why they should be regarded as aids in 
the criticism of the text of Zephaniah, and not rather as worthy 
monuments of ancient interpretation. Whether they vindicate 
their value for criticism or not, they can help to fix the history 
of the Hebrew text only to the time when the earliest of them 
was made. Beyond this point, if the text obtained does not 
commend itself as a true copy of the autograph, external criti- 
cism by the help of translations must yield to Conjectural Criti- 
cism. A tree only the top of which is visible above some obstruc- 
tion illustrates quite accurately what can be known of the text of 
Zephaniah. The angles of convergence must indicate where the 
continuation of the trunk is, and where branches and trunk join. 
The present investigation thus resolves itself into a test of the 
Hebrew transmission at three points, the exact location of which 
is unknown. This somewhat anticipatory statement has, it is 
hoped, outlined with sufficient clearness the general trend of the 
discussion and vindicated the method employed. 

IV. The little that the Hebrew text in editions and manu- 
scripts offers may be at once presented. 1 I 1 rvpm R. JTpSn , cf. 
Peshitta; pDK R. j'DK, due to the accidental joining of the 
strokes for i and final j . I 4 iKtf K. (3 MSS.) Dtf, cf. Septua- 
gint. run R. Kim, error due to the forgetfulness of a scribe 
who carried his copy in his memory from clause to clause ; n 

i Kittel's text is used as a basis ; B. = Baer and Delitzsch ; G. =Ginsburg (not his latest 
edition of 1908); T. = Thiele; W. = Walton's Polyglot; M. = Massoretic Notes; R. = De 
Rossi's Collations ; K. = Kennicott's Collations as cited by R. 

6 The Text of Zephaniah. 

R. n&O , to avoid possible confusion due to asyndeton. I 6 r\UJ 
R. rnu, error of vision. I 6 i#p:j G. B. wpi . I 8 ontfn by R. 
ontyn SD Sy , error of memory, cf. I 4 . I 12 r\^3 R. DV3 , cf. Septu- 
agint. I 16 D^y R. D^n , error of hearing, frequent with gut- 
turals. 2 1 itfKnpnn B. nstehpnn. 2 2 p;j T. W. -pnD; D-\DU 
clause (3) omitted, R. (6 MSS.), K. (8 MSS.), homoioteleuton. 
2 4 niEhr B. mtfnr . 2 7 onntf M. orratf (G. does not point this 
word). 2 9 S U R. DTI, error of memory, cf. I 4 ; OITT W. DN3\ 
2 12 '3in R. mn, cf. Peshitta. 2 14 np_ B. np ; ^D3 M. ^D3. 
2 16 1T M. iiyKi, odd expression, occurring here only, changed to the 
usual one. 3 1 runiD G. B. ntnn . 3 2 x 1 ? R. xbi , cf. I 4 ; ^K 
R. SKI, cf. I 4 . 3 4 D^ma T. o-mia. 3 9 naj; 1 ? R. na^Si , cf. I 4 . 
3 10 ^ia rn omitted, R. (IMS.), K. (1 MS.), cf. Septuagint and 
Peshitta. 3 14 T?jn B. 'iSjji . 3 15 p-K R. ^K, error of memory, 
cf. I 4 ; <Kvn M. R. K. ^-in; ;n B. G. ^. 3 18 ^y M. R. 
!j^^ , decision must be arbitrary, cf. Peshitta and the Revised Ver- 
sion. 3 20 D^ry 1 ? R. M. orrr/7, cf. 3 18 . 

The printed texts from Walton to Kittel are identical except 
in a few pointings and matres lectionis. The sporadic readings 
in the collations are either due to the versions or are explainable as 
common corruptions in manuscript transmission. Other explana- 
tions than those given above may be equally satisfactory ; but the 
true reading is nowhere in doubt, as each variant has the support 
of only a few manuscripts at most. It is evident that the arche- 
type of the manuscripts and printed texts here represented has 
been transmitted with remarkable accuracy. 



1. The history of Jerome's translation may be divided into 
three epochs of unequal length, the first and second each cul- 
minating in an important recension of the text, the third con- 
tinuing into the present. The first period is one of conflict 
between it and the Old Latin which it was meant to supersede. 
The new translation met with violent opposition from many 
quarters, and its introduction was therefore very gradual. The 

The Vulgate. 7 

fact that the older version persisted and the method by which a 
text had to be transmitted conspired together to rob Jerome's 
translation of its purity in this conflict of almost four centuries. 
It could conquer the older version only by absorbing many of its 
characteristics, while every copy that was made both transmitted 
and increased errors. The power of the Church was being more 
and more concentrated and its influence so extended that it was 
gradually becoming the dominant force in Western Europe ; but 
the authority of the^ Bible, which was the foundation on which 
the whole structure of the ecclesiastical hierarchy was felt to rest, 
was being dissipated more and more, because hardly two copies 
of it were in agreement. A supreme papacy needed an official 
text, and it remained for Charlemagne, who was actuated mainly 
by liturgical motives, to establish one by means of the recension 
undertaken at his behest by Alcuin. Theodulf (f 821) made an 
independent recension at about the same time. The Vulgate 
which was thus established doubtless differed in many important 
particulars from Jerome's autograph, but unfortunately very little 
is known of the history of the text during these centuries of con- 
flict; and the students of Latin Bible texts are consequently 
unable with any degree of fulness to trace out the process by 
which the Carolingian Vulgate was evolved. The verses quoted 
by the church fathers of the period and the few incidental 
remarks scattered here and there through the pages of their 
writings throw but a feeble light into the darkness, which begins 
to lift only in the last century (VIII). 

II. Toward the close of this epoch and in the next the Vul- 
gate takes higher and higher rank. Wherever the Church goes, 
it goes as the official version of the Word of God, while Latin 
becomes everywhere the language of worship. The artificial 
unity of language thus established was a powerful factor in build- 
ing up an ecclesiastical sovereignty that practically obliterated 
national boundaries. The Hildebrandian Papacy had been all 
but impossible without the Vulgate, which had for many centuries, 
first through use in missionary propaganda, and then in the litur- 
gies and lectionaries of worship, been welding together the 
diverse elements of which it was composed. The torch of learn- 
ing, though it burned most dimly, was borne along by the Church 

8 The Text of Zephaniah. 

alone during this dark period; and the only text-book in most 
curricula was the Vulgate. Copies were multiplied with great 
rapidity in the schools and monasteries. Again, as was inevita- 
ble, the text became so corrupt that many recensions were made. 
These sporadic attempts could, however, bring about no perma- 
nent improvement, because manuscripts were so widely distributed 
that concerted effort was impossible, while the scribes' choice of 
exemplars to copy was controlled by the flimsiest critical princi- 
ples, if by any. 1 Even the early printed editions were for the 
most part set up from such manuscripts as were near at hand. It 
remained for the Council of Trent to suggest the remedy that 
the discovery of printing had made possible, and the Sixtine- 
Clementine edition is the result of a decree passed by that body. 
With but few exceptions the 8000 extant manuscripts of the 
Vulgate belong to this period. The text to be found in them is 
almost uniformly corrupt. 

111. The third period of the Vulgate's history begins with 
the Clementine text of 1592. It is still the official text of the 
Vatican. Many reprints of it have been made ; but no edition, 
embracing the results of the latest discoveries and based on 
approved critical methods, has yet appeared, at least not for the 
Old Testament. Heyse and Tischendorf's pretentious Biblia 
Sacra Lat. Vet. Test. Hieronymo interprete (1873) is practically 
identical with Bagster's cheap reprint. 2 The Latin column of 
Funk and Wagnall's popular Hexapla Bible (1906) differs only 
in a few punctuations from the de luxe edition, Biblia Sacra Vul- 
gata (Critice edidit P. M. Helzenover, 1906), in which at least 
one misprint has escaped the proofreader. 3 Vercellone's Sacra 
Vulgatae Editionis Sixti V et Clementis VIII (Rome, 1861) is 
generally regarded as the best. 4 

1 Cf. Berger, Histoire de la Vulgate pendant les premiers siecles du moyen dge, 
Paris, 1893, pp. 329, 330. 

2 The differences between them in Zephaniah are as follows, Bagster's text being the 
first cited; I 3 " 5 Coelicaeli, 2 2<a Domini Dm, 2 6 speciosamSpeciosam, 3 14 lauda, 
lauda, jubila, jubila, corde, corde. 3 1 ' nolitimere noli timer e. 3 ig fuerat,fuerat. 

3 3" Adijcies for adjicies. 

* Now and again more or less extensive excursions have been made into the field of 
the textual criticism of the Vulgate Old Testament, but on the whole it is still an unex- 
plored domain. Berger, in the introduction to the work already mentioned, gives a very 
satisfactory sketch of what has been accomplished both in the Old and New Testaments. 
The book itself takes rank as a classic in Vulgate studies and contains a complete bibli- 

The Vulgate. 9 

IV. The Vulgate manuscripts must be considered in their 
geographical distribution; for three main types of texts, kept 
more or less distinct from each other by natural boundaries, are 
clearly defined. Ireland and Spain because of their location both 
remained for the most part isolated from the rest of Europe. 
The Vulgate text, which was early taken to these countries, was 
thus kept separated from the main continental current of trans- 
mission. As the purity of a text is, generally speaking, inversely 
proportioned to the number of times it has been copied, the rate 
of corruption of manuscripts was much less rapid in Ireland and 
Spain than elsewhere. But Irish missionaries and Irish monks 
kept carrying the Irish text to different parts of the continent ; 
and in the first year of the ninth century the Alcuin recension 
brought the Irish type of text back into the main stream of trans- 
mission, for he is known to have sent to York for manuscripts to 
be used in his work. 1 Theodulf seems to have been familiar with 
the manuscripts in use in the South of France, and his collations 
may have brought into the main current many characteristic 
Spanish readings. In the Clementine text these three types are 
blended, for manuscripts from many places were collated for it. 
As compared with each other, the pure Irish type is much better 
than the pure Spanish. The known national characteristics of 
the two peoples lead to the inference that Irish manuscripts 
would be less ornamental and more accurate, and this is confirmed 
by all that is known of the types. 

V. It is clear from what has been said that a comparison of 
manuscripts of these three types will yield the earliest obtainable 
text. The Codex Amiatinus is earlier than the Alcuin recension, 2 
and the Codex Toletanus antedates Theodulf. 3 For the conti- 
nental type, in lieu of anything better, the Clementine must needs 
be used. The results of such a comparison for Zephaniah are as 
follows: 4 I 1 Sophoniam filium Chusi. A. Sofoniam filium Cusi 

1 Jaffe, Monumenta Alcuiniana, p. 346. 

3 A very interesting account of how the age of this, the best of the Irish manuscripts, 
was finally fixed is to be found in Studio, Biblica et Ecclesiastica, Oxford, 1890, II, pp. 
273 ff. 

3 A description of these manuscripts may be found in Berger's Histoire de la Vulgate^ 
etc., pp. 37 f. and pp. 12 f. 

* The Clementine text is used as a basis. A=Amiatinus; T.=Toletanus. The collation 
of A. is taken from Heyse and Tischendorf s apparatus ; that of T. from Migne's Patrolo- 
ffia Latina, XXIX, p. 1027. Italics have been used to indicate the readings which 
deserve the preference. Where more definite criteria fail (cf. 2 14 ), It is necessary, since 
relative values have not yet been fixed, to decide by simple majority rule. Readings 
that are evidently corruptions have been marked as such. 

10 The Text of Zephaniah. 

(Jerome is known to have aspirated the Begadkefat ; cf . Lagarde's 
Onomastica, index), filii Godaliae A. T. filium Godaliae (this 
is perhaps an Old Latin reading as it agrees with the Septuagint). 
ftlii Amariae filii T. filium Amariae filium (the sense demands 
the genitive). Hheciae A. Ezechiae (p was not aspirated by 
Jerome in transliteration ; cf . Lagarde's Onomastica, index) . Amon 
A. Ammon. Judae T. Juda. I 3 volatilia A. T. volatile. I 6 
super omnem A. omnem. omnem .... qui ingreditur T. 
omnes .... qui ingrediuntur. I 11 Pilae T. filiae (corruption). 
disperierunt T. dispergerunt (corruption). I^faecibus A. feci- 
bus (spelled foecibus, Jer. 48"; the spelling fex is allowable; cf. 
Harper ' Latin Dictionary , p. 744). faciet A. faciat (corruption). 
I 14 Juxta est A. Juxta et (corruption). I 17 corpora A. corpus. 
I 18 faciet cunctis T. faciet Dominus cunctis (interpretative addi- 
tion, suggested perhaps by I 12 ). 2 2 super vos ira A. ira; ante- 
quam clause (2) omitted T. (This may be Old Latin, cf. p. 31.) 
Indignationis A. furoris (in the Liber de Divinis Scripturis 
sive Speculum, XVI, De Libro Sophoniae, this same variant 
occurs in an evident Vulgate text, and therefore the reading of 
A. is to be adopted). 2 s qui T. quia (corruption). 2 B Philisthino- 
rum A. Philistinorum (cf. I 1 ), inhabitator T. habitator. 2' 
pecorum T. ovium (this may be Old Latin). 2 7 remanserit A. 
manserit. 2 H quae T. qui (corruption). 2 9 Gomorrha A. Gomorra 
(cf. I 1 ), in aeternum T. in sempiternum (this may be another 
Old Latin reading), eoset . . . . illos A. T. illos . . . . illos (the 
agreement of A. and T. is hard to explain unless they represent 

the Old Latin ; the Septuagint has avroi>s K<U avrovs, thus 

the agreement with it is only partial). 2 11 viri A. T. vir (in a 
quotation, evidently made from memory, Augustine has vir with 
adorabit. He seems to have changed the number of the verb to 
turn this Hebraism into intelligible Latin, whereas the Clemen- 
tine text has changed the number of the noun). 2 12 et vos Aethi- 
opes T. et vos et Aethiopes (dittography). 2 13 Speciosam A. 
T. Speciosa (Jerome's translation of Nineveh is hardly intelligible 
in Latin, and the unusual fern. sing. adj. was early corrupted into 
the ordinary neut. plu.). 2 14 quoniam T. quum. 2 15 civitas glori- 
osa A. gloriosa civitas (accidental inversion). 3 2 confisa T. 
confixa (corruption), appropinquavit A. adpropiavit (corrup- 

The Vulgate. 11 

tion). 3 5 mane mane A. T. mane (homoioteleutonic omission; 
or perhaps better, the Hebraism was early removed), lucem A. 
luce (corruption). 3 6 disperdidi A. disperdi, T. disperdit (cor- 
ruptions), neque ullo A. nee ullo. 3 7 dixi attamen A. dixit 
tamen (corruption), suscipies T. suscipe (as timebis was read, 
suscipe must be a corruption). 3 s et effundam A. T. ut effun- 
dam (the reading ut may be accepted, not only because it is sup- 
ported by these two ancient manuscripts, but because it brings 
out the meaning of the Hebrew better; per se a corruption is 
possible either way ; the Old Latin has et). indignationem T. 
oninem indignationem (dittography due to following omnem). 
3 9 invocent A. T. vocent. 3 13 mendacium et non T. menda- 
cium non (accidental omission). 3 H Jubila A. Jubilate (inter- 
pretative with Israel in distributive sense). 3 17 salvabit T. 
salvabit te (perhaps due to Old Latin influence; cf. Septuagint). 
exsultabit T. et exsultabit (cf. Septuagint; more likely, how- 
ever, an ordinary sporadic reading). 3 19 earn quae ejecta fuerat 
T. ea quae electa fuerant (corruption). 3 20 tempore quo con- 
gregabo T. tempore congregabo (monograph y). 

VI. Since the distance of the text now established from the 
autograph must still be measured in centuries, many Old Latin 
elements that crept in after Jerome had finished his work may be 
contained in it. The Spanish text as a whole is known to betray 
an especially strong Old Latin influence, and perhaps the syno- 
nyms of T. in 2 7> 8 , as well as other readings peculiar to this man- 
uscript (3 17 ), come from this source. The Old Latin of Zephaniah 
has not survived, 1 and consequently it cannot be directly deter- 
mined how much of it, if anything, has passed into the Vulgate 
either originally through Jerome himself, who sometimes con- 
sciously, and perhaps more often unconsciously, incorporated its 
readings, or through subsequent confusions due to their transmis- 
sion side by side. In the belief that they would be of interest, 
and, perhaps, even of importance in this connection, a collection 
of quotations from the early Latin Fathers was made. 3 It was 

1 There seems to be a manuscript in the Vatican which contains the last eight verses of 
the Old Latin of Zephaniah ; cf. Swete, Introduction to the Old Testament in Greek, 
p. 97. 

2 After the collection was completed it was found that a similar collection had already 
been published ; cf. Journal of Theological Studies, 1903, p. 76. The results of these 
two independent examinations are in substantial agreement. 

12 The Text of Zephaniah. 

rather disappointing to find that only a few of the Latin ecclesi- 
astical writers before the middle of the fifth century were cited 
in the critical editions of their works as having referred to 
Zephaniah. In Tertullian only an allusion to the dies irae was 
to be found. A single clause occurs in Nolan us: 

l llb JExterminati sunt omnes qui exultati fuerant auro et 

Vulgate: disperierunt omnes involuti argento. This can be 
regarded only as an expansive allusion to Zephaniah. Cassian 
quotes a clause, the thought of which is of such a nature that 
divergence in its expression is practically impossible except in 
particles : 

I 12b Qui dicunt in cordibus suis, non faciet Dominus bene, sed 
neque faciet male. 

Vulgate: Qui dicunt in cordibus suis : non faciet bene domi- 
nus, et non faciet male. 

More than a third of the book can be recovered from Cyprian, 
Augustine and Tyconius. 1 For the purposes of comparison that 
which seems to be genuine Old Latin has been here placed 
between the Vulgate and the Septuagint. 

1 The Liber de Divinis Scripturis sive Speculum is here regarded as the work of 
Augustine, to whom it is attributed by its editor for the Corpus Scriptorum Ecclesiasti- 
corum Latinorum, Vienna Academy. It is, however, by many attributed to an unknown 
author. Augustine's capriciousness in quotation is abundantly sustained. His text 
agrees with that of the Vulgate in five passages, I4h.7a- nb t 2 i-3, ssa-ia-isa. F Or 21.3 and 312 
he has also quoted the Old Latin. His two quotations of 2 11 are so mingled that he 
must have quoted from memory in both cases. 

Augustine (1). Praevalebit dominus adversus eos et exterminabit omnes deos gen- 
tium terrae, et adorabunt eum unus quisque de loco suo, omnes insulae gentium. 

Augustine (2). Horribilis Dominus super eos, et exterminabit omnes deos terrae, et 
adorabit eum vir de loco suo, omnes insulae gentium. 

Vulgate. Horribilis Dominus super eos, et attenuabit omnes deos terrae; et adora- 
bunt eum vir de loco suo, omnes insulae Gentium. 

The Vulgate. 



(I 8 ' 3 ) Congregans congre- 
gabo omnia a facie terrae, 
dicit Dominus : Congregans 
hoininem, et pecus, congre- 
gans volatile eoeli, et pisces 

marls : et disperdam 

homines a facie terrae 

(P) Silete a facie Domini Dei : 
quia juxta est dies Domini 
quia praeparavit Dominus 
hostiam, sanctificavit voca- 
tos suos. (l llb ) Disperierunt 
omnes i n v o 1 u t i argento. 
(list- i4a) Aedificabunt domos, 
etnon habitabunt: et planta- 
bunt vineas, et non bibent 
vinum earum. Juxta est dies 
Domini magnus. (I 14b - 16 ) Vox 
die! Domini amara, tribula- 
bitur ibi fortis. Dies irae 
dies ilia, dies tribulationis 
et angustiae, dies calamitatis 
et miseriae, dies tenebrarum 
et caliginis, dies nebulae et 
turbinis, dies tubae et clan- 
goris super civitates munitas, 
et super angulos excelsos. 
(Ii7b- isa) Et effundetur san- 
guis eorum sicut humus, et 
corpora eorum sicut stercora. 
Sed et argentum eorum, et 
aurum eorum non poterit 
liberare eos in die irae Dom- 
ini. (2i-8) Convenite, congre- 
gamini gens non amabilis: 
Priusquam pariat jussio quasi 
pulverem transeuntem diem, 

antequam veniat 

super vos dies furoris Dom- 
ini. Quaerite Dominum, 


(!, Cyprian) Defectlone 
deticiat a facie terrae dicit 
Dominus, deficiat homo et 
pecudes, deficiant volucres 
caeli et pisces marls et au- 
feram iniquos a facie terrae. 
(V, Cyprian) Metuite a facie 
Domini Dei, quoniam prope 
est dies ejus; quia paravit 
Dominus sacriflcium suum, 
sanctificavit vocatos suos. 
(l llb , Speculum) Disperierunt 
omnes qui exaltantur in 
argento [et auro]. (lb. ", 
Cyprian) Aedificabunt domos 
et non inhabitabunt, et insti- 
tuent vineas et non bibent 
vinum earum, quia prope est 
dies Domini, (li^-is, Specu- 
lum) Vox diei domini amara 
et dura constituta, dies po- 
tens, dies iracundiae dies ille, 
dies tribulationis et necessi- 
tatis, dies infelicitatis et ex- 
terminii, dies tenebrarum et 
tempestatis, dies nubis et cali- 
ginis, dies tubae et clamoris 
super civitates firmas et super 
angulos excelsos. (I 17b> 18a , 
Speculum) Et effundam san- 
guinem eorum sicut limum, 
et carnes eorum sicut stercus 1 
bourn et argentum et aurum 
eorum non poterit liberare 
eos in die irae domini. 
(2i-3, Speculum) Convenite 
et congregamini populus in- 
disciplinatus, priusquam 
emciamini sicut flos prae- 
teriens priusquam super- 
veniat super vos dies iracun- 
diae domini. Quaerite dom- 


1 stercora in another place. 

dirb Trpoff&irov rijs 7775, 
Kvpios. 'ExXiTT^Tw avdp(t)iros 

Kal KTT?)Vr) ^XlTT^TW TO. 7TC- 

Teivd TOV ovpavov Kal ol 
TTJS da\d<ro"rjs ' . . . . Kal 
TOVS dv6fwvs dirb irpo- 
XT)? yijs .... (!') 
atrb irpoffdirov 
TOU 0eou ' 5i6rt ^771)5 17 
-rj/ji^pa TOU Kup/ou, 8ri ^r 

Kf TOI>S K\r}TOVS O.VTOJ. (l llb ) 

.... (i}\o0pti0Trjffa,v Trdvres ol 
dpyvpiy. (Ii3b.i4a) 

Kal ov /j. 
avrais ' Kal KaTa<pVTev<rov<riv 
d/LtTreXtDvas, Kal ov ^ irluffi. 
TOV olvov auTwv. "Ort ^771)$ 17 
Wpa Kvpiov .... (Ii4b.i6) 
(fxavTj ij^pas Kvpiov iriKpa Kal 
ffK\-rjpa rtraKTai. Avvarij 
6pyi)S, i) 
^X^ews Kal 
awplas Kal d<pavifffJU)v, 
GKOTOVS Kal yv6<pov, 
ve0A7;s Kal oplx^*}*, 
(T(X7ri77os Kal Kpav- 
7^s lirl ras 7r6Xets TCLS 6xv- 
/oas, Kal irl ras ywvlas ras 

%ee? rb a?/xa avr&v 

Kal rds ffdpKas avruv a>s 

/86Xj3tra. Kai rb dpyvpiov 

avruv Kal rb xP vff tov avrGiv 

ov IJ.T] dtivyrai %\t(rdai av~ 

TOVS tv ijiJ-tpa dpyijs Kvpiov. 

(2 1 * 8 ) Svvdx^re, 

6-rjTe rb fdvos rb 

irpb TOV yevtffdat v/ 

Trpb TOV 

yfdpav dv/Jiov Kvpiov. 

(Tare Tbv Kvpiov irdvTes Tairei- 


The Text of Zephaniah. 


omnes mansueti terrae, qui 
judicium ejus estis operati: 
quaerite just urn, quaerite 
mansuetum : si quomodo ab- 
scondamini in die furoris 
Domini. (2i3_35a)Et extendet 
manum suam super Aquilo- 

nem, et ponet Specio- 

sam in solitudinem, et in 
invium, et quasi desertum. 
Et accubabunt in medio ejus 
greges, omnes bestiae Gen- 
tium : et onocrotalus, et eri- 
cius in liminibus ejus mora- 
buntur : vox cantantis in 
fenestra, corvus in superlimi- 
n a r i, quoniam attenuabo 
robur ejus. Haec est civitas 
gloriosa habitans in confiden- 
tia: quae dicebat in corde 
suo: Ego sum, et extra me 
nonest aliaamplius: quomodo 
facta est in desertum cubile 
bestiae? omnis, qui transit 
per earn, sibilabit, et move- 
bit manum suam. Vae pro- 
vocatrix, et redempta civitas, 
columba. Nonaudivitvocem, 
et non suscepit disciplinam : 
in Domino non est confisa, 
ad Deum suum non appropin- 
quavit. Principes ejus in 
medio ejus quasi leones rugi- 
entes : judices ejus lupi ves- 
pere, non relinquebant in 
mane. Prophetae ej us vesani , 
viri infideles : sacerdotes ejus 
polluerunt sanctum, injuste 
egerunt contra legem. Domi- 
nus Justus in medio ejus non 
faciet iniquitatem. (38) 
expecta me, dicit Dominus, 
in die resurrectionis meae in 
futurum, quia judicium meum 
ut congregem Gentes et colli- 
gam regna : ut effundam su- 
per eos indignationem meam, 


inum omnes humiles terrae, 
aequitatem operamini, et 
justitiam quaerite, et respon- 
dete ea , ut protegamini in die 
irae domini. (2i3-35a ) Tycon- 
ius) Et extendet manum suam 
in Aquilonem et ponet illam 
Nineve exterminium sine 
aqua in desertum, et pascen- 
tur in medio ejus greges 
omnes bestiae terrae. et 
chameleontes, et hericii in 
laquearibus ejus cubabunt, et 
bestiae vocem dabunt in fos- 
sis ejus, et corvi in partis ejus 
quoniam cedrus altitude ejus. 
Civitas contemnens quae 
habitat in spe, quae dicit in 
corde suo Ego sum, et non est 
post me adhuc ! Quomodo 
facta est in exterminium pas- 
cua bestiarum ! Omnis qui 
transit per illam sibilabit, et 
movebit manus suas. O in- 
lustris et redempta civitas, 
columba quae non audit vo- 
cem, non recepit disciplinam. 
in Domino non est connsa, et 
ad Deum suum non adpro- 
pinquavit, principes ejus in 
ea ut leones frementes, judices 
ejus ut lupi Arabiae non re- 
linquebant in mane, profetae 
ejus spiritu elati viri contemp- 
tores, sacerdotes ejus profa- 
nant sacra e t conscelerant 
legem. Dominus autem Justus 
in medio ejus, non faciet in- 
justum. 1 


vol 7775, Kplfjia e"p7ife<r0e, 

1 Cyprian's exegesis of 3 1 ' 2 
shows the substantial agree- 
ment of his text with that 
ofTyconius : C o 1 u m b a 
non exaudit vocem, id 
est, praeclara et redempta 
civitas non recipit doctrinam 
et in Dominum fidens non 
fuit. In the Speculum a 
clause of 3 4 is quoted: Sacer- 
dotes ejus contaminant sancta 
et reprobant legem. This is 
perhaps a quotation from 
memory, as Tyconius has a 
reputation for accuracy, es- 
pecially in long passages. 

diroKpivecrde avrd, oirus <r/ce- 
iracrQiJTe ev 77/^/39 6/37775 Ku- 
plov. (2 13 3 5a ) Kal eKTevei TT> 
Xeipa avrov eirl fiopbav .... 
Kal #77<rei TTJV Nti/eu^ els 
d(pat>i<r fJJbv avvdpov, ws eprj/MV. 
Kal ve^ffovraL ev jtteVy avrijs 
TTofyma, Kal travra ra Orjpla 
TTJS 7775, Kal x a / 



tv TOIS 

rb avd<TTT}fj.a 
77 7r6Xis 77 <pav\iffTpia, 77 
KaroiKOVffa tir" 1 ATT/SI, 77 
\tyovffa tv KapStq. ai/TTjs, 
'E7w ftfu, Kal OVK ftrri ner" 1 
ifj^ en * TTcDs tyevfiOf) eis 
a<t>avLffpJbv, vo/J-rj dfiplwv; iras 
6 diairopev6/j,vos Si 1 avrrjs <rv- 
ptet, Kal Kivfoei ras xetpas 
avrov. "ft 77 eirt<pav7]S Kal 
diro\\VTpufJi.t}>r) 7r6Xts, 77 TTC- 
purrepa OVK elff^Kovffe <p<i)i>r)S ' 
O$K fS^aro iratSelav, iri rf 
Kvplfp OVK eTrcTrot^et, Kai irpbs 
rbv Qebv avrijs OVK tfyyicrev. 
01 apxovres avrijs ws \VKOL 
TIJS 'Apa/3/as, ovx vireKlirovTO 
els TO irput 01 irpotpTjrat 
avTijs irvVfj.aTo<f>6poi, avdpes 
Kara(ppovT]Tai ' iepeis avrrjs 
f3ef3-r)\ov<ri rd ayia, Kal dffe- 
(Bovo't vbfjkov. '0 d Kvpios 
OLKCLLOS v jn&rtjj auTT^s, Kal ov 

'l aOLKOV ' (3 8 ) . . . 

^7ei Kfyuos, ds 
dvaffrdcreibs pov els 
ftapTvpiov ' 5t6 rb Kpl/ma /JLOV els 
o-vvaywyds 0vu>v, TOV elff- 
dej-a<r6ai j3a<rtXe?s, TOV 

The Vulgate. 15 


..... (3- 13a ) Quia tune red- (8, Cyprian) Expecta me, & afoote ira<rav dpyijv 6vfwv 

dam populis labium electum, dicit Dominus, in die resur- /gg.iaax "Q Tt 

ut vocent omnes in nomine rectionis meae in testimon- 

Domini, et serviant ei humero ium ; quoniam j u d i c i u m *f^W* tirl Xaoi/j 

uno. Ultra flumina .... de- meum ad congregationes gen- ffo-v ets yeveav ayrT/s, TOV iri- 

ferent munus mini. In die tium, ut excipiam reges et /taXear^at Train-as TO 6vofML 

ilia non confunderis super effundam super eos iram K . - 5ouXej j etJ/ aljT /j 

cunctis adinventionibus tuis, meam. (3-is, Augustine) ' 

, , 

quibus praevaricata es in me : Transvertam in populos lin- U7rd &"* v ? " a - E/c 7re / ) a TW " 

quia tune auferam de medio guam et progenies ej us, ut in- iroTa/j-iov 'At0w7r/as 6i<rov<ri 

tui magniloquos superbiae vocent omnes nomen Domini 0v<rlas (ju>i. 'Ej> 

tuae et non adjicies exaltari et ser-viant ei sub jugo uno ; a ^ 
amplius in montesancto meo. finibus flummum Aethiopae , 

Et derelinquam in medio tui adferenthostiasmihi. Inillo 7ra " TW J' 

populum pauperem, et ege- die confunderis ex omnibus ffov i & v fyrtpyffas ets fyt 

num : et sperabunt in nomine adinventionibus tuis, quas 3rt rbre irepieXu airb <rov rd 

Domini. Reliquiae Israel. . . inpie egisti in me ; quia tune 0auX Z(ruoTa rm CSecis aov 
auferam abs te pravitates in- n ~ 

juriae tuae ; et jam non ad- ** OVK Ti ** TOV 

jicies, ut magnificeris super fJya\avxi)<rai tirl rb 6pos 
montem sanctum meum, et S^tbv /*ou. Kai 
subrelinquam in te populum 

mansuetum et humilem ; et 

verebantur a nomine Domini, "*", "* 
qui reliqui fuerint Israel. ^T^ TOU 6v6fj.affTos Kupiou Oi 

KaraXouroi TOU 

In these verses positive proof of Jerome's use of the Old Latin 
is not to be found. There are a few agreements, but these may 
well be accidental. 1 The remarkable differences, even in places 
where greater similarity would hardly have been surprising 
because of the nature of the ideas to be expressed, seem to pre- 
clude literary dependence on Jerome's part; for this could be 
established only by more striking agreements in more character- 
istic passages. The so-called Itala Question does not present 
itself in connection with these quotations. In only one case 
(3 1 * 2 ) are the same verses recovered from two sources. In one of 
these it is in an interpretation and not in a quotation, and this 
may well account for the slight differences found. It may now be 
stated positively that the text already established must be con- 
sidered as the purest text of the Vulgate of Zephaniah that can 
be obtained. 

1 Cf. I 7 sanctificavit vocatos suos; I 18 angulos excelsos; 2 1 convenite, congregamini; 
3* in Domini non est confisa, (et) ad Deum suum non (ap) adpropinquavit. 

16 The Text of Zephaniah. 

VII. No more emphatic proof of the high esteem in which 
the Vulgate is still held could be offered than the fact that 
modern Catholic scholarship is about to engage in the stupendous 
task of a new revision which will, when completed, be the crown- 
ing tribute of Latin Christianity to St. Jerome. 1 Doubtless the 
choiceness of its diction and the majesty of its style have been 
largely instrumental in raising this version to the commanding 
position which it has so long occupied in the Catholic Church ; 
but it could not continue to usurp the place of the inspired 
Hebrew Old Testament so entirely, if its general faithfulness as a 
translation were not beyond dispute. In Textual Criticism, 
however, accuracy in detail is the measure of a version's value; 
and entire consistency in translation, even to the complete subor- 
dination of all matters of style and diction, is the translator's 
chief virtue. The Latin text of Zephaniah reveals frequent con- 
flict beween the careful translator and the literary artist. Occa- 
sionally Jerome's faithfulness to the Hebrew leads him to do vio- 
lence to the Latin idiom (l u congregans congregabo ; 2 11 adora- 
bunt eum vir de loco suo). More frequently he is satisfied with 
an ad sensum rendering from which the reading of his exemplar 
could never be recovered without the help of the Massoretic Text 
(2 T * qui remanserit de domo Juda=mMV JV3 rriKiy; ibi= orp 1 ?;; ; 
3* injuste egerunt cowra=iDon; 3 6 dum non est qui transeat= 
"131 y ll ?3D; non remanente viro, neque ullo habitatore=r#ft ETK sl ?3D 
atfr ; 3 7 omnia, in quibus visitavi eam=rrS;rrnpJD "i^xSD; 3 17 
fortis^ ipse salvabit=yw 113J). His translations of participles 
prove him a firm believer in the principle of varietas delectat 
(participle = participle I 4 ; participle with article = participle I 12 ; 
participle with article = relative clause I 12 ; participle = relative 
clause 3 8 ; participle = adjective 3 s ; participle = noun I 18 ; parti- 
ciple = finite independent verb I 14 ; cf. further 2 14 , where finite 
independent verb = participle, and I 4 , 1", 2 5 - 6 , 3 8 , where 3tfr is in 
each case differently rendered). Connectives he supplies or 
omits quite arbitrarily (I 11 , I 18 , 2 1 , 3 5 ' 6 , 3 9 ), and occasionally he 
inserts the copula (1 s , 2 10 , 2 15 eveniet). Prepositions are for the 

The work is to be directed by the Rt. Rev. F. A. Gasquet, Abbot President of the 
English Benedictines. The many uncatalogued cathedral libraries of Spain and Italy 
are being systematically overhauled, and special copies of the Clementine text are to be 
printed to aid in the work of collation. 

The Vulgate. 17 

sake of variety or interpretatively supplied, omitted or changed 
(2 a , 2% 3 3 , 3 7 , 3 16 ; in I 8 - 4 he seems to distinguish between Sj?D and 
|D, the former being rendered by ab, the latter by de). He 
sometimes shows a very accurate knowledge of Hebrew syntax 

(I 8 et erit visitabo^mpQi rrni , the Septuagint has 

Kai rrcH /cat cKSiKijo-w ; 3 7 diluculo surgentes corruperunt= 

irrniyn iD'Diyn). In matters of vocabulary he is, however, not a 
safe guide. riDT destroy and StfJ pollute were unknown to him. 
Speciosam in 2 13 is due rather to his failure to understand the 
passage than to his fondness for translating proper names (cf. l n , 
Pilae) ; at any rate his etymology of Nineveh, if he read the 
word, is far-fetched. 1 The richness of his Latin vocabulary is of 
course largely responsible for his lack of consistency in the choice 
of words. For almost every Hebrew word to be translated there 
were many Latin equivalents and near-equivalents at his command. 
pS is rendered in the Vulgate Old Testament by morari (2 14 ), com- 
morari, demorari, man ere, remanere, permanere, quiescere, requi- 
escere, habitare, esse, resider effing ere tentoria, dormire (cf. further 
* 49 13 , 59 16 , Job 27 7 , II Sam. 12 16 for less accurate or mistaken ren- 
derings). *\D33=porrigere, par are, concupiscere, desiderium esse, 
amabilis (2 1 ). Within Zephaniah the same root is sometimes trans- 
lated by different words (2 16 , 3", 3 14 ; 3 7 , 3 11 ). In I 3 - 4 'ivon is ren- 
dered by disperdam ; and mjj in I 11 is very properly rendered by dis- 
perire, which is the regular passive of disperdere ; 3 in 2 5 , however, 
disperdere is the translation of T3KH, which in 2 IS is rendered by 
perdere, and in 2 7 mDJ is translated by perire. Pertinent illus- 
trations might be multiplied almost indefinitely, but enough have 
been given to indicate Jerome's general habit of translation and 
to show how wide are the limits within which the equation, Vul- 
gate equals Massoretic Text, may with entire safety be allowed 
to obtain. The readings that demand more special consideration 
will be noted later. For the rest of the text it can be shown on 
the basis of the above analysis either that the present Hebrew 
and the Vulgate agree, or that proof of their disagreement is 

His derivation of Nineveh is perhaps based on some Midrashic interpretation. He 
has connected mrj with n*O or HU, cf. Jer. 6" and Zeph. 2. 
* Cf. Harper's Latin Dictionary (Lewis and Short), p. 592. 

18 The Text of Zephaniah. 



I. There is no apparatus criticus for the study of the 
Peshitta text of Zephaniah, and with the exception of Ceriani's 
photolithographic reproduction of the Cod. Ambrosianus no man- 
uscripts are available. As far as can be gathered from the 
scattered and incidental notices of various writers, there are only 
a few old Syriac manuscripts containing this book in the libraries 
of Europe. That there are none in Berlin rests on the authority 
of Strack. In England those earlier than the seventeenth century 
are British Museum Add. 14,432, 14,443 and 14,468 (I 1 ' 6 ) ; Cam- 
bridge L. e. 2.4, Uni. Add. 1965, Buchanan Bible. In lieu of 
manuscripts the printed editions, of which there are five, must be 
used to establish a critical apparatus for the text. Of these the 
Syriac text of the Paris Polyglot is the earliest (1645). This 
was reproduced in Walton's London Polyglot (1657), and again, 
but without vowels, by Lee for the British Bible Society (1821). 
A Syriac Bible was printed in Nestorian characters and with 
Nestorian vowels by American missionaries in Urmiah (1852). 
More recently the Dominicans of Mosul have printed a text 
(1887-1892). It is difficult to determine the critical value of 
these editions. P. (=Paris Polyglot), W. (= Walton's Polyglot) 
and L. (=Lee's text), are generally allowed to count as only one 
witness, because their differences are either misprints or improve- 
ments in spelling. That L. was used for U. (=Urmiah) can be 
deduced from the text itself. 1 It has not as yet been made certain 
whether M. (=Mosul) has independent value or not, because those 
competent to judge seem to have been unable to obtain copies. 2 
The text of P. is known to have been taken from the manuscript 
Syriaque 6 of the JBibliotheque Nationale, which dates from the 
seventeenth century. As a manuscript it has no special merit, 

1 Cf. Nestle, Hastings' Dictionary of the Bible, IV, p. 651 a . 

2 Cf. Barnes, An Apparatus Criticus to Chronicles in the Peshitta Version, with a 
Discussion of the Value of the Codex Ambrosianus, Introduction. 

The Peshitta. 19 

and it seems to have been used only because it was convenient for 
the printers to handle. The sixth tome of Walton's Polyglot 
(pp. 19 if.) contains a collation of two manuscripts, Usher and 
Pocock. In the Prolegomena to this work (p. 165 2 ) it is stated 
that Us. ( = Usher) was copied "from a codex of the Patriarch 
of Antioch," who is the head of the Maronites. In all likelihood 
this was an old codex not on sale. According to Barnes (Journal 
of Theological Studies, II, p. 186), Lee had access to the Bu- 
chanan Bible and to Cambridge L. e. 2.4, and it may therefore be 
assumed that he did not find in them any readings which seemed 
to warrant a departure from the London Polyglot. These as well 
as Cod. Ambrosianus are Jacobite manuscripts. If manuscripts 
were used for U. and M., they were undoubtedly of Nestorian 
and Jacobite or Maronite character respectively. 1 These few 
facts and probabilities, in which practically all that is known 
about the origin of these texts is comprised, can in themselves 
hardly support any positive conclusions ; but in the light of the 
history of Syrian Christianity they are of paramount importance 
for the textual criticism of the Peshitta. 

II. The Peshitta version owes its survival largely to the 
Christological heresies of the fifth century. After the Council 
of Ephesus (431) the followers of Nestorius were so bitterly 
persecuted by their Monophysitic opponents that the heresy 
taught by him was speedily stamped out in Italy and Greece. 
The Oriental Nestorians, over whom the ecclesiastical control of 
Rome and Constantinople was but feeble because they were sep- 
arated both by language and character from the Christians of 
Europe, maintained their peculiar tenets despite all opposition, 
and Syria became virtually a theological battle-ground. The 
Monophy sites were victorious because of the powerful advocacy 
of Anastasius and Zeno, and they succeeded in driving the Nes- 
torians more deeply into the territory of the Sassanian kings of 
Persia. Since the Gospel was first preached within their king- 
dom, these kings had watched with suspicion the "aliens who 
had embraced the religion, and who might favor the cause, of 
the hereditary foes of their realm;" 2 but now that they were 

1 Rahlfs made the assumption that Nestorian manuscripts were used by the American 
missionaries (Beltrage zur Textkritik der Peschita, Z.A.T.W., 1889, pp. 161 ff.). 
a Cf. Gibbon, Decline and Fall of the Roman Empire, XLVII. 

20 The Text of Zephaniah. 

rebels against the Roman Empire and fugitives from Roman 
jurisdiction, they were eagerly welcomed, and in the year 483 or 
484 1 at the Synod of Beth Lapat Nestorianism was officially 
adopted as its confession of faith by the Christian Church in 
Persia. The Monophysites were themselves anathematized by 
the Council of Chalcedon (451), and the same influence of lan- 
guage and character operated to perpetuate this heresy in Syria. 
They were in turn harassed by the Nestorians, for whom apparent 
defeat had become a triumph through the powerful allies gained, 
and their ecclesiastical organization was accomplished only with 
the greatest difficulty by Jacob Baradaeus, from whom they 
obtained the name of Jacobites. The odium theologicum thus 
kindled between Nestorians and Jacobites has never ceased to 
burn. Toward the close of the following century many of the 
Syrian Christians who had escaped both Nestorianism and Mono- 
physitism and who were called Melkites because of their loyalty 
to the Empire, were wrecked on the rock of Monothelitism ; and 
a third sect resulted whose members are called Maronites. They 
never became entirely free from Roman influence and were finally 
brought back into the Church, when certain minor concessions of 
ritual and clerical privilege were made by the Papacy. 

III. The Peshitta remained the official version of Scripture 
for these three sects ; and though Arabic or Persian became 
their vernacular after the Mohammedan conquest, the Bible con- 
tinued to be read in the sacred language. Their common accept- 
ance of the Peshitta in spite of their lasting hostility to each 
other amounts to proof positive that the Peshitta antedates the 
schisms which separated them; and the schisms, in that they 
would tend to produce three distinct lines of transmission, give 
to Textual Criticism its only means of determining an ancient text. 

(A.) Where all the authorities agree, it may be safely affirmed 
that the text is older than the last quarter of the fifth century. 

(B.) Am. (^Ambrosianus) and Us. together establish the 
West-Syrian reading, for one is Jacobite and the other Maronite. 

(C.) If U. contains any distinctly Nestorian readings, they 
ought to be easily recognized because they stand alone. 

1 Cf. Noldeke, Aufsatze zur persischen Geschichte, p. 107. 

The Peshitta. 21 

(D.) In the absence of more positive criteria U. may be 
allowed to decide between West-Syrian readings. These vaguely 
general and by no means absolute rules, 1 aided here and there by 
the scholia of Bar Hebraeus and the quotations of other writers, 
must in the absence of anything better fix this important text for 
the entire fifteen centuries or more of its existence. Though 
Assemani, himself a Syrian, has written a tome of 950 pages 
concerning Syrian Christianity, 2 he throws little light on the 
history of the Peshitta as such ; and little more is now known of 
its origin 3 than Theodore of Mopsuestia seems to have known 
when he wrote : 

fjpfjirivcvTai 8e ravra ets fJ^v rrjv r<av Svpon/ Trap' OTOV S^TTOTC, ovBe yap 
lyvaKTTai /u-e'xpt TT/S riy/xepov ocrns -TTOTC OVTOS eoTtv. 4 

IV. The following is a collation of P., W., Us., Po. ( = 
Pococl$), U., M. and Am. with L. : I 8 * ^ * * ^\ Am. 

I 9 * .cov^Jf^o p o> .oouaJj^ao . I 11 * . ^V ft ^ Ain "" ^" * A . I 11 * 

Am. *oZ ? . I 12 P Am. Us.* P?. I 15 * l?co ?? _Am. fecofo. I 15 * 
|J^I P. P^l. I 17 * U 1*1 -Am. UJ] ^1^. I 18 * gold and 
silver Am. silver and gold. 2 2 * Pr^o(3) U. M. Pr^ . 2 9 - 10 - 13 - 14 - 16 * 
"^-4.1^0-.] U". M. omit both alephs, Am. omits the first. 2 9 * U^ 
Am. tt^X 2 n * ^X^l U. ^X 1 !?. 2 12 Am. v cJJ additional. 
(3 7 - 8 break in Am.) 3 11 ^-r- Am. U.* 
U. M. *-^. 3 17 * h^r^ M. ^r^. 3 19 
-M. U. Us. Am. * v ociZZoi^ ? \i] oi^as. 320 jy[ jj. 
Us. Am. add at the beginning of the verse * ^a-^-l 001 U^^ ouo . 
3' V OOUJL^^_U. M. Am. * v ooi-J.^ . 

The readings to be preferred according to the rules formulated 
above have been starred. With one exception the variants are 
of no importance, consisting either in omissions and additions of 
, ?, and ^, or in differences of spelling. In 3 19 - 20 the collation gives 
a reading which commends itself as original. The text obtained 
from these different lines of transmission contains inner-Syraic 
corruptions, and these must therefore be very early. In I 9 Po. 

1 The rules here formulated agree substantially with those given by Rahlfs (Z.A.T.W., 
1889, pp. 161-210) , though much less positively stated. 

2 Assemani, Bibliolheca Orientalis, IV. 

3 Cf. Berg, The Influence of the Septuagint upon the Peshitta Psalter, New York, 1895. 
Cf. Migne, Patrologia Graeca, LXVI, p. 241. 

22 The Text of Zephaniah. 

has corrected one of these by reading sf 01 ^i^ for ^pqi * i, SP . The 

others are l^*- for i^^ (2 11 ) ; IH^ (pointed Ij^ in W.) for 1-=^ 
(2 14 , cf. Brockelmanrf s Lexicon Syriacum, p. 258 b , and Ez. 
17 3.-^. ^ , for )^oi (3 ? c f. i"). 

V. Bar Hebraeus cites Zephaniah in the following verses, 
quoting at most a clause though generally only a word: 1 I 1 , I 2 , 
I 3 - 3 - 3 - 3 , I 8 , I 10 ' 10 , I 11 - 11 , I 17 , 2 7 - 7 , 2 12 , 2 13 , 2 14 - 14 ," 2 15 , 3 4 , 3 5 , 3 9 , 3 15 . In 
I 11 one of the three codices collated by Moritz agrees with Am. 
in omitting the final o of cjoZ? . J n 2 7 , where the editions all have 
] *v>i^> ? Bar Hebraeus seems to have read i-^- i-aurs (i n ripa 
maris). This may be an explanation of the geographical location 
of Askalon; some connection with the S^n of 2 5 - 6 ' 7 is not unlikely. 
The remainder of his citations agree with the text of the editions. 
The scholia have no textual value, being either on the vocaliza- 
tion of words or of an interpretative character. Quotations of 
Zephaniah must be exceedingly rare in early Syriac religious 
literature, because a patient search of many indices and footnotes 
yielded only a few allusions to Zephaniah by Ephraem Syrus in 
his poetical Homilies, and two partial quotations of the same verse 
(3 9 ) by Aphraates, in which he does not differ from the accepted 
text. 2 

VI. The translation of Zephaniah, while literal, is not 
slavish, and its style is smooth and flowing. The similarity of 
Hebrew and Syriac in idiom and vocabulary was evidently of 
great help to the translator ; but still the Peshitta, as the Vulgate, 
falls far short of that accuracy of detail and consistency in 
translation which gives a version its chief value for Textual 
Criticism. The data which show the general character of the 
translation, and which thus, though of little or no importance per se, 
indicate where possible variants may be looked for and where not, 
may be at once collected and dismissed from further considera- 
tion. 3 

1 Cf. Moritz, Oregorii Bar Hebraei in Duodecim Prophetas Minores Scholia, 
Leipzig, 1882. 

2 It was impossible to find out whether the recension of Jacob of Edessa made in 704-5 
was still extant; cf. Swete, Introduction to the Old Testament in Greek, p. 116, n. 4; and 
Ryssel, Untersuchungen uber die Textgestalt und die Echtheit des Buches Micha, 
p. 173. 

s Cf. Introduction; the Syriac readings are always mentioned first. The plus and 
minus of the Peshitta in regard to Vau are not noted; it is added about forty times and 
not once omitted. 

/ Of THE 


V or 

'""fa* Peshitta. 23 

I* | *1 *1^ ma (always except in Ez.). I 4 ? additional (name of 
the Chemarim with the priests = name of the Chemarim with that 
of the priests). ! 6 Va (2) additional. T Part. = part, and looi 
(cf. l a , 3 5 , part.^impf.). l 6 ^ (2) additional. I 6 ^ ^<* = :DJ 
nnKD (cf. Is. 59 13 ). I 7 Ur* 1r^=ni7V ^1N (only in the Minor 
Prophets and Ez.). I 7 0=0 (cf. 3 20 , few). I 7 ^iol=anpn (cf. 
Jer. 12 3 ). I 8 Part.=part. and looi w ith ? (cf. 2 15 ). I 11 ^^o^= 
(of. 3 1 , <Jo-) ^ ^ *' additional. l ia U^tt 1 ?!. I 14 

^r^ eoi UU-D J. i V>> 1-05 oi^oo* ooi uOu^O=VtJFI mn s DV D'lp 
a^ip (the changes here are for the sake of clearness). 
1" 1M=ni3^ (cf. 1 18 =1>^05). I 17 Impf. with ?=i consecutive 
with perf. I 18 Gold and silver = silver and gold (cf. Am.). I 18 
jj^ajoo ^a^? ^nSnaj ]K n^D (this is perhaps a good interpretation, 
but not a very exact translation, cf. I 9 ll"^ ^aok- ^offiXs ^ 
jnsDn^^nn). I 18 Part.=impf. 2 4 vor^P ?i-*'l 3 =ninnE;K . 2 6 
Part, with ^ part, construct. 2 9 ^^ i-^-^i? 5^=^ Vn. 
2 9 looLi additional. 2 9 ' 5 ^-lr ia -'i^^ additional. 2 11 ^o^o^^nSK (for 
theological reasons). 2 12 minus suffix and HDH . 2 14 l^aiQ^j Ueu^ 
= s u in^n . 2 14 J?i its houses =in its capitals. 2 14 01Q -^ j^na . 2 16 
jua^j? ^SnSizziaijy SD . 2 15 oioJ and i^opo additional (due to the 
fact that the following verses were referred to Nineveh). 3 1 
I&JL^ additional (interpretative). 3 1 ^a*-=n:rn (this is read as 
the name of the prophet because of the interpretation just men- 
tioned [2 16 J; the order of the words is changed for the same 
reason). 3 2 ? additional. 3 5 V ^ 1 3a ^= l ?ij; jnr *6i . 3 6 AS^4 
=1D27J. 3 6 Part, with ?=part. 3 6 -^> ^o=" l ?aD = pKO (cf. 2 6 ). 

3 7 ?=!jK . 3 7 ^^^ additional (interpretative). 3 7 j^j^cVa nty ^3. 

3 8 p] ^0^05 ^Dip. 3 8 ^^ additional (this verb is inserted to 
guard against ambiguity). 3 8 h-^=Djfl=a|H (cf. I 16 ). 3 9 ^-r-<no 
= m^3 (cf. 3 n =T^i-<n). 3 9 ? withimpf.=inf. of purpose (2). 
3 10 Shall bring to me offerings shall bring my offerings. 3 11 

aipD (cf. 3 17 , ^a-^ = ^a")pa). 3 
(cf. 2 7 ). 3 15 **u* additional (cf. 3 7 ). 3 16 

= |ry . 3" Vr* ^^=^2^ "^^ 3 18 
!j 1 3j;D SD (it is unnecessary to suppose that D^D was read). 
3 20 ooi additional (cf . I 14 ' 16 ). 3 20 Impf. =inf . with suffix. 3 20 ? l^ 
with part, and pronoun subject^inf. with suffix. 

24 The Text of Zephaniah. 

It is evident from this collection of "peculiarities" that the 
motive of the Peshitta translator was religious rather than 
scholarly, and that he desired to make a readable rather than an 
exact translation. He much preferred expansion to condensation. 
Interpretative additions, especially in places where the style of 
the Hebrew is concise or elliptical, are not infrequent (2 15 , 3 1 , 3 7 , 
3 8 ). There is a marked preference for long sentences, and these 
are formed by adding connective particles (passim). Pronouns 
are both supplied and omitted (2 12 , 3 11 ). Interpretations and 
paraphrases are occasionally found (I 4 , I 9 , I 18 ). There is at least 
one change for theological reasons (2 11 ). A word denoting a 
general conception is sometimes substituted for one that denotes 
a particular part of the conception (2 14 ). A plural is often used 
to render a collective (I 3 , 2 14 ). Minor changes of order, the 
reason for which is not clear, also occur (I 18 , 3 8 ). Gross ignorance 
of Hebrew syntax and vocabulary cannot be laid to his charge 
(cf., however, 1", 3 1 ). Ryssel's general estimate of the Peshitta 
of Micah 1 will serve equally well for that of Zephaniah. His 
words are : Fassen wir . . . unser Urtheil tlber den Syrer zusam- 
men, so muss die grosse formelle Gewandtheid anerkannt werden, 
mit welcher er die Gedanken des hebr. Textes ins Syrische 
tiberzutragen versteht, und der leichte, fltissige Stil, in dem alle 
Unebenheiten des Ausdrucks beseitigt sind; dabei schreibt er 
korrekt und vermeidet deshalb meist Hebraismen. 



I. For many centuries after its origin the Septuagint was 
a potent religious force, first among Hellenized Jews and later 
more especially among Christians. Its importance is shown by 
the translations of Aquila, Symmachus and Theodotian indirectly, 2 
and directly by Origen's Hexapla and the recensions of Hesychius 
and Lucian. Through the gradual ascendancy of Rome, its place 

1 Ryssel, Untersuchungen uber die Textgestalt und die Echtheitdes Buches Micha, 
p. 171. 

It is generally agreed that these translations were made in antagonism either to the 
Septuagint or to each other. 

The Septuagint. 25 

was, however, more and more usurped by the Vulgate, and its 
direct religious influence continued only in the many secondary 
versions that were based on it. 1 During this time through the 
mixture and conflation of recensions and translations a process 
which was much facilitated by the lazy and ignorant use of 
Origen's Hexaplaric Septuagint the text of the manuscripts be- 
came exceedingly corrupt. Humanism in its passionate love for 
the literature of Occidental antiquity, and the Protestant Reforma- 
tion in that it rejected the authority of the Vulgate while its 
formal principle demanded an authoritative Scripture, combined to 
revive a critical interest in the Septuagint which has been steadily 
growing; but it has long ceased to undergo recension for religious 
motives, and the printing-press has checked all further corruption 
by eclectic manuscript transmission. The history of the Septua- 
gint thus falls into two general epochs, which may be called the 
Epoch of Construction and the Epoch of Reconstruction. Between 
these lies the period of manuscript transmission in which the 
second epoch must find its material with which to work. Many 
editions of the Septuagint have appeared, but the process of 
reconstruction is still far from complete. 2 The great problem is 
to recover the pre-Hexaplaric (pre-Origenic) text; but this can 
be obtained only after the Hexaplaric, Hesychian and Lucianic 
texts, which lie confused together in the manuscripts, have been 
separated from each other. The three types thus obtained would, 
after the recensional elements of each had been removed, represent 
the texts current in Palestine, Egypt and Syria in the early and 
late third century, and their collation would yield a very early 
Greek text. Along these lines comparatively little has as yet 
been done. 3 The extant Hexaplaric fragments have been collected 

1 In the East, where it is still recited by the\Orthodox Church in the Ecclesiastical 
Offices, it lost much of its influence over the thought and life of the people. Swete 
Introduction to the Old Testament in Greek, p. 433. 

2 As many as sixty-three editions and reprints between the Complutensian text and 
that of the larger Cambridge Septuagint (now in preparation) are enumerated by Nestle 
and Swete. The editions from which the reprints have been made are he Complutensian 
(4), the Aldine (6), the Sixtine (45) and the Alexandrian [Grabian] (5). The Cambridge 
Manual Septuagint completes a total of sixty-five. There are also several facsimile and 
photolithographic editions of manuscripts, but these are not generally accessible. Many 
editions of single books or groups of books have appeared; the text of Zephaniah seems 
never to have been separately published. 

3 The larger Cambridge Septuagint will when completed be valuable mainly for its 
critical apparatus, for in its text it will but repeat the Manual Septuagint text of Codex B. 

26 The Text of Zephaniah. 

by Field in his Hexapla Origenis, but the text is not restored in 
a connected form. Lagarde began the reconstruction of a pro- 
visional Lucianic text, but only one volume of his work appeared 
before his death. 1 The Hesychian recension has not yet been so 
much as definitely identified. 2 The collations of H. P. (= Vetus 
Testamentum G-raecum, cum variis Lectionibtis, ed. Robertus 
Holmes [ . . . editionem a Roberto Holmes inchoatarn continuavit 
Jacobus Parsons], Oxford, 1798-1827), and S. (= Cambridge 
Manual Septuagint, The Old Testament in Greek according to 
the Septuagint, H. B. Swete, 1887-1894, 2 1895-1899) contain 
practically all the evidence of manuscripts and editions ; but the 
former has been severely criticised on the score both of accuracy 
and arrangement, 3 while the latter contains the variations of only 
the important uncials. 

II. The pre-Hexaplaric text of Zephaniah cannot thus be 
directly and positively established; indirectly something may, 
however, be done. In the apparatus criticus resulting from the 
combination of H. P. and S. there are numerous itacistic and 
sporadic readings which are easily recognized as such. Many 
rival readings, having good manuscript support, in regard to 
which nothing positive can be determined because of their nature, 
are also to be found; but since Cod. B. (=Vaticanus) on the 
whole presents the version in its oldest form, the balance of 
probability is in its favor in these cases. There are, however, 
several pronounced variations from B. and its supporters B. of 

1 Cf. Lagarde, Symmicta, II, pp. 137-148. 

2 To what extent the Hesychian recension is still accessible in manuscripts and 
versions of the Septuagint is uncertain Swete. Field made no mention of it in the 
Introduction to his Hexapla, although he discussed Lucian and his work extensively. 
Ceriani made the claim that the Codex Marchalianus (Q., XII) of the Prophets agrees 
very closely with the text presupposed in the Egyptian versions and in the works of 
Cyril of Alexandria, and that it is supported by 26, 109, 198 and 306. According to Tischen- 
dorf this codex belongs to the recension of Eusebius and Pamphilus, i. e. , it Is Hexa- 
plaric. The Hesychian group in Ezekiel according to Cornill is 49, 68, 87, 90, 91, 228, 238. 
Lagarde and Cornill thought that this recension was to be sought in the Aldina edition, 
which generally follows 68 even in its mistakes; but Stekhoven claimed that the Complu- 
tensian text in the Minor Prophets agrees with 40, a manuscript which is closely related 
to the text used by Cyril of Alexandria and therefore to Hesychius. Grabe found the 
recension in Codex B. For the remaining books of the Old Testament (i. e., with the 
exception of the Prophets) we have as yet no published list of manuscripts containing 
a probable Hesychian text Swete. 

3 A complete stemma exhibiting the filiations of these manuscripts and recensions 
cannot be made from the collations of H. P. Moore, Judges, p. 14. 

The Septuagint. 27 

course represents all manuscripts not cited as differing from it, at 
least in so far as the collations are dependable, which are of such 
a character that either they or the readings of B. from which 
they differ must be due to the recensions ; and for the attribution 
of at least some of these more or less positive criteria are available. 
The critical notes in the margin and text of the Syro-Hexapla in 
some cases indicate a choice, and in others a reading is shown to 
be due to Lucian by the known characteristics of his work. The 
fragments of Origen's Hexapla (fifth column) collected by Field, 
the Syro-Hexaplaric version and the Old Latin fragments can also 
here and there be used as a test. The text of B., thus confirmed 
or corrected as the case may be by the available evidence, may be 
accepted as original. Although many elements of uncertainty 
must remain in a text thus established, these will be in matters of 
detail which are of importance mainly for the editor of a critical 
edition, and which do not materially affect the use of the version 
for Old Testament Textual Criticism. 

111. The Greek manuscripts of Zephaniah contained in the 
collations are the following: 


A. (III.) Alexandrinus. V. 

K. Sinaiticus. IV. 

Q. (XII.) Marchalianus. VI. 

V. (23.) Venetus. VIII. 

T. Cryptoferratensis. IX. 

22. British Museum Reg. 1, B. 2. XII. 

26. Vat. Gr. 556. XI. 

36. Vat. Gr. 347. XIII. 

40. Dorotheus Moldaviens. XII. 

42. Demetrius Moldaviens. XII. 

49. Laur. XI. 4. XI. 

51. Laur. X. 8. XI. 

62. Ox. New Coll. XIII. 

68. St. Mark's, Gr. 5. XV. 

86. Barber V. 45. X. 

87. Chigi 2. IX. 
91. Vat. Ottob. Gr. 452. XI. 
95. Vindobon, Th. Gr. 163. ? 
97. Vat. Gr. 1153. X. 

106. Bibl. Comm. Gr. 187, Ferrara. XV. 

114. Evora, Carthus. 2. ? 

28 The Text of Zephaniah. 


131. 1 Vindobon, Th. Gr. 23. XII 

147. Ox., Bodl. Laur. 30. ? 

153. Vat. Gr. 273. X. 

185. Vindobon, Th. Gr. 18. XI. 

198. Paris, Nat. Gr. 14. IX. 

228. Vat. Gr. 1764. XIII. 

233. Vat. Gr. 2067. XII. 

238. Vat. Gr. 1153. IX. 

239. St. Salvator Bonon. 641. XI. 

240. Laur. VI. 22. XIII. 

310. Mosq. Syn. 209. XI. 

311. Mosq. Syn. 341. XI. 

IV. Approximately 500 different readings are noted in H. P. 
and S. (a) To this total K* has contributed a large number. 
He seems to have been a very poor copyist, as the following 
specimens of his work will show: I 3 yx^ves (t^vcs), I 4 ^tpa (^et/oa), 
I 15 ras TrdXts, 2 10 TravTO/cpdro/aav, 3 3 lire\L(f>Or)O-av. (>) Evident cor- 
ruptions of all kinds abound: I 12 dya&OTronJo-ei (dya^OTroi^o-r;), I 14 
Taxwv) and raxvvr) (ra^eux), 2 4 SieaTracr/xoo^and Biecnrap/^evr) 
2 5 K/OITWI/ (KpryTwv), 2 7 KaraXvTrots (KaraXotVots) , 3 8 e^e'Xcn 
3 6 KarcWa (/carcWao-a). (c) The sporadic readings of single or of 
related manuscripts are numerous : spelling, 'lov&W ('lovSa) ; mood, 
I 7 vXa/?eicr0at (evXa/?er0e) ; tense, 1 s eKXeiTrera) (eKXiTreVa)) ; number, 
3 B avraij/ (avr^s), I 3 eKXtTreroxrav (e/cXiTreTco) ; person 2 1 i)/xas (v/xas), 
3 6 e^cpTyttaxrav (c^epry/xwo-w) ; case, I 6 Sw/xcuri (Sw/xara) ; preposition, I 10 
OTTO (cTrt'), I 10 eKKevTowTwv (aTTo/cevTowTcov) , 2 2 eX^etv (cTTcX^etv) ; syno- 
nyms, I 16 tcr^vpds (o^vpds), dSt/cuxs and dvo/xoxs (dcreySetixs) ; words of 
similar appearance, 3 12 TroXw (TT/DCIVI/), 3 6 tixfrOrjcrav (rj^avLcrO^a-av) , 
3" Trpovxy? (TT/OOO-^S); additions, I 4 ev (before Jerusalem), 3 2 <rov; 
omissions, I 1 6s: homoioteleutonic, 2 9 /xe'w? to /xeviy; dittography, 

I 4 7Tt *IOV&XV Kttt 7Ti *Iov8aV (7Tt 'lovStt Kttt'). 

V. Between a large number of rival readings which both have 
good manuscript support, decision must, as has already been sug- 
gested, be arbitrary. I 4 icpoiv tepeW. I 7 ^roe/xaKe ^Toi/xa<re. I 10 
v rrj T7/Apa CKUVT) ev Kf.cvy ry -jy/xe/oa. I 11 Oprjvrja-aTe. OpyvtiTt. I 13 
ov xi^| KaroiKr](TOv<rw ov /x^ Karot/CT/crowrtv. I 14 ^ ^xupa ly/xepa. I 18 
([i/Xov iyXovs. 2 4 'AaKaXcov 'AcTKaXcov <rrat. 2 8 

1 130 and 131 is the same manuscript. It is by Lagarde assigned to the thirteenth 
century (cf. Z.A.T.W., 1908, p. 11). 238 is said to be a copy of 87. 

The Septuagint. 29 

2 11 7ri^>av7/o-CTat em^av^s efrrat (cf. Joel 2 11 ' 31 , Hab. I 7 ). 3 a OVK 
ovSe. 3 4 01 icpcis tepcis. 3 6 &o8eveo-0ai, SioSeveiv. 3 8 8ta Trapa. It 
is with equivalents of which these are representative that the 
elusive Hesychian recension may sometime be connected, unless 
indeed the view that it was a new version now lost ultimately 
prevail. 1 Comparatively few variants remain, after B's readings 
have been accepted in all the cases that belong to this class. 

VI. Lucian had a double purpose in revising the Septuagint 
text. He wished to improve its Greek and at the same time make it 
conform more nearly to the original. His reverence for the Sep- 
tuagint sometimes led him to place two translations side by side. 
In supplying lacunae he made use of the translations of Aquila, 
Symmachus, and Theodotian. His text also has interpolations that 
serve only to indicate the nexus of the thought or to make an obscure 
passage clearer. He seems to have allowed himself to introduce 
only minor changes for the sake of better Greek. An occasional 
removal of stiffness by a slight change of construction, and the sub- 
stitution of a singular for a plural predicate with a neuter subject, 
of a more familar word or form for one less familar, of one com- 
pound verb for another, and of a simple for a compound verb or 
vice versa, as far as now known, mark the extent of his literary 
revision. It is evident that there are no absolute criteria for 
detecting his merely literary changes, and therefore many variants 
of which one or the other is perhaps due to him belong to the 
class of which illustrations have already been given. ( V.) Cor- 
rections according to the Hebrew and interpretative additions 
may, however, be identified with more or less certainty. Accord- 
ing to Stekhoven the following readings are Lucianic: l a iravra 
additional; I 3 TO. trKavSaAa o-vv rots dcre/Jecrtv; I 4 TO>V )8aa\t/w., /u-era ran/ 
tepetov additional; I 6 Kara TOV Mc'X^o/x; I 12 TOVS Aeyovras; I 17 KXD; 
2 a ^/xcpas additional ; 2 s ^T^crare SiKawxrw^v ^TT/crare irpaor-qra Kat 

vra; 2 13 ve/XTycrcrat ; 2 15 eycvcro; 3 s \\vTpo)fJivrj ; 3 1 * 
3 20 ov additional. To these may be added: I 1 eyeVero; 
I 18 Svv^rai; 2 13 KTvu>, fjioV) airoXu*, 6rf<T<^\ 3 4 eis TOV vo/u,ov; 3 7 ctTrov, 
8u<0apT<u; 3 13 ov /txij. There are also two readings from the other 
Greek versions which may have been introduced by Lucian: I 16 

Aquila; 3 8 eye/oo-e'ws pov cuwvuxs, Symmachus. These 

Nova potius versio quam ' Septuagintae ' interpretum dicendaGr&be. 

30 The Text of Zephaniah. 

readings, none of which are to be found in B., must all be rejected 
as recensional; they give no indication as to the nature of Lucian's 
Hebrew text. 

VII. (a). In its text and margin the Syro-Hexapla has a few 
important critical notes: I 4 */xra TWV lepeW. I 5 * *at TOV? Trpoo-- 
(margin) ; 2 2 * yuepa (note ; hoc ex reliquis) , * Trpo TOV 
<' v/xas ^/xepav 6vfj.ov Kvptou; 2 7 -j-r^s ^aXa(r<r>ys, -r- 'lovSa 1 ; 
3 6 * cts <ws Kat OVK aTreKpvftfj Kat OVK eyvto doW'av cv aTratTTyo'et (margin) ; 
3 6 -r-KaTO"7ra<ra VTrep^aVovs fi<f>avi(T0r)<raVy 3 10 * TrpocrSe^o^tat ei/ 8te<T7rap- 
/xe'vots fte (margin) ; 3 14 0vyarep; -4-Aeyet Kvptos. (5) From the text 
of the Syro-Hexapla additional data may be gathered: I 7 Kat 
omitted before ^yta/ce; I 9 rt Travras additional, cov omitted; I 11 
Kat before tguXoOpevOrjo-av omitted; I 16 the order of words is O-KOTOVS 
Kat yvo<ov; 2 2 Ovpov additional; 2 14 Kat before KopaKcs omitted; 3 6 
the order is Kpt/ua avrov Swo-et, VIKOS for VCIKOS; 3 8 Tr)v opyirjv fjtov -jrao-av 
opyyv Ovfjiov /xov was read ; vTroXeti^o/xat for vTroXiy^o/xat. (c) Field's 
fragments of the fifth column of the Hexapla, in so far as they 
were not derived from the Syro-Hexapla, were obtained from 
Codd. 86 and Q, and from the commentaries of Jerome, Theo- 
dore of Mopsuestia and Cyril of Alexandria on the Minor 
Prophets. Some have thus already been cited ; the rest are here 
added. 1 s Kat aaOevrjo-ov&iv ot dcreySers; I 4 Kat ra ovo/xara TWV tepcwv; 
I 5 Kara TOV /Sao-iXews; I 9 Kat KStK>/<ra> fjL<f>av<*)<s ?rt ra TrpoVvAa; I 10 dwro 
7rvAr;s aTTOKCvrowTcov ; I 11 ot KarotKovvres rrjv KaraKCKO/x/Aeviyv ; I 12 e^epev- 
v^o*a); I 16 OXtyetvs, dwptas Kat d^>avto-p,ov) ; I 18 o-vvreXctav Kat 
2 1 (rvvd^drjTe. Kat o*w8e'^>yTe; 2 3 Kpt/xa, Kat aTroKptVeo^e avra; 2 4 
eo-rat; 2 B TrdpotKOt Kpryrtuv; 2 6 Kp^rry; 2 9 Kat Aa/xao^Kos, w? 

2 14 ws ^a/xatAcovTCS ; 3 1 w 17 e7rt<^av^s Kat a7ro\e\VTp(afievr) TroAts, 
^ 7Tpto*Tepa; 3 3 XvKOt T^5 'ApayStas ; 3 6 ev 8ta^>^opa, ycovtat avraiv; 3 8 cts 
fjfjiepOLV dvao"Tao"co>s p,ov cts fAaprvpiov ; 3 9 on TOT p,eTaoTpi^<o CTTI Xaov? 
yXo>o"o*av fits yeveav avr^s; 3 10 CK Treparwv Trora/xoiv 'At^tOTrtas otcrovo-t 
^vo-tas ;u,ot; 3 18 ws ev ^epa eopr^s, ovat. 

VIII. In the passages represented in these collections the 
text of B. is for the most part confirmed; and its readings, except 

1 In the course of transmission an obelus has evidently fallen out before 
The one before 'Iotf5a is perhaps due to the fact that in some manuscripts a new line was 
begun with this word, for the diacritical marks were repeated before the first word of a 
new line. 

The Septuagint. 31 

such as are about to be individually considered, may be at once 
adopted in preference to their alternates. 

1 B All the evidence goes to show that /cat TOVS Trpotr/cvvowras 
was absent from the original text of the Septuagint, and these 
words must be deleted from B. 

I 9 In omitting CTTI Travras B. seems to have no better support 
than 40 and 239. Field and the Syro-Hexaplaric text disagree. 
It is necessary to insert this in B. The Syro-Hexaplaric omission 
of @cov is not explained by a note, but that this word was in the 
original Septuagint is attested by the Vulgate. 

2 2 The last clause is asterisked in the Syro-Hexapla. In the 
preceding clause OvpSv is added with opyrfv (^Kjnn), and in the 
clause asterisked o/oyJ/s seems to have been read for Ovpov. There 
seems to have been some confusion between these clauses the 
initial words of which are the same. At least "aberant igitur 
haec a ' Septuaginta ' " does not at once follow, especially as the 
Old Latin preserved in the Speculum omits the second of these 
clauses and retains the third. The same omission is suggested by 
a corrector of Cod. Sinaiticus ( c - b ). These clauses are peculiarly 
liable to omission by homoioteleuton, as 233, Cod. Toletanus and 
several Hebrew manuscripts demonstrate. In view of this fact, 
and more especially because of the evident confusion, it seems 
unnecessary to delete either one clause or the other. 

2 7 The Syro-Hexapla misrepresents Origen in 'suggesting that 
his fifth column read airo TT/OOO-WTTOU wuiv -f- 'lov&x X . The obelus 
must be placed before the first word. These words were perhaps 
incorporated into the text by someone who did not understand 
the absolute use of /caraAveiv in the sense of to lodge. 

3 5 ' 6 According to the Syro-Hexaplaric notes and text Origen's 
fifth column read: 

Kpifjua. avrov Swcret * eis <<*>? /cat OVK a.7TKpvf3rj /cat OVK cyvw (<rav) dSt/ctav 
ev aTTonrrjcru X /cat OVK ets vt/cos dSt/ctav ev 8uL<f>06pa -r- /caT(T7ra<7a VTrepr;- 
<dVovs ^avto-^o-av X . The signs are again misplaced, for the 
last three words are certainly not a Septuagint addition. It is 
known that Origen sometimes gave .two readimgs where the 
Septuagint differs widely from the Hebrew, and that he then 
indicated the Hebrew current in his time by an asterisk and the 
Septuagint by an obelus. Evidently someone who was ignorant 

32 The Text of Zephaniah. 

of this special method of indicating a doublette has arbitrarily 
brought about conformity with the general practice. One of the 
metobeli must be deleted, and the other must be substituted for 
the obelus. An obelus must be placed before the first ev. The 
Septuagint reading thus obtained makes fairly good sense, but it 
cannot be regarded as an attempted translation of the Hebrew 
that has come down to us. There are indications in the collation 
of H. P. that vet/cos must be read for vt/<os and Sia<d/oafor Sia<0opa. 
The Syriac for eV aTramJo-ei is Ik^sks. Field seems to have read 
this as ]&o A^ and the meaning may be in doubt. The Origenic 
reading thus becomes a triple gloss, "in doubt," "and not in 
dispute dSi/a'av," "in disagreement". The trouble seems to have 
been due to the words ntzto SlK, the first of which was so translated 
that the second had no apparent government. Comments occa- 
sioned by this supplanted the text, while a slight change in one 
case gave a good sense, and in another a possible translation of 
the Hebrew word (n&2i = ev 8ta<f>06pa) . The original Septuagint 
for n^ is thus unknown, but ev aTratTT/o-et /cat OVK eis vet/cos dStKiav h 
8ia<f>06pa must be deleted. 

3 8 D;N is represented in the Syro-Hexapla, but according to Field 
it was not represented in Origen's fifth column. It is absent from 
the Old Latin. As 0v/x,os and 0/0717 translate ^x, pin, DJN, rn3p and 
non indifferently, it was impossible to translate literally where 
three of these words occur together without repeating one or the 
other of them. One Greek word thus sometimes represents two 
Hebrew words (cf. Is. 13 13 , Jer. 4 8 , Zeph. 2 a ), and this may well 
be the case here. 

3 10 The clause 7rpocr8e'o/>uu iv Steo-Trap/xeVoi? /xov is to be deleted 
because it is absent from the Hexaplaric text and the Old Latin 
of the Speculum. 

3 12 The reading vTroAij^o/xat is an early inner-Greek corruption 
for i}7ro\i\l/ofjicu. 

The text of the Cambridge Manual Septuagint, which is based 
on a facsimile edition, is to be preferred to that of H. P., which 
represents a copy of B. made by the Dutch Septuagint editor, 
L. Bos. Where the Manual contains T instead of B. (3 9b -3 20 ), 
H. P. is to be preferred, and so xxu ^o-o/xat (3 19 ) and on (with 
, 3 20 ) are to be inserted. oAe'0/>evo- (2 11 ) should be eoAo- 

The Septuagint. 33 

0/oevo-e (cf. 3 7 ). iraiSiav (3 a ) should be TraiScutv (cf. 3 7 ). aXwvos (2 9 ) 
must be corrected to dA.os (Putamus dA.os inter pretatos, id est^ sails ; 
sed ab imperitis, qui 0i/x,o>vtav, Aoc es, acervum, frumenti vel 
frugum, putaverunt, pro oAos additis duabus litteris, o> et v, quasi 
ad consequential^, frugum, aXwvos, Aoc es, areae positum Jerome) . 
As there seem to be no quotations of Zephaniah in the early 
Greek Church Fathers, the New Testament Apocrypha, the New 
Testament, Josephus, Philo or the Old Testament Apocrypha, no 
earlier text than that now established can be obtained. 

IX. The readings of the Septuagint which illustrate the 
general character of the translation without proving differences 
of text can now be presented. From these it will be seen that 
the Septuagint has no general characteristics which it does not 
share with the Peshitta or the Vulgate, or with both. 1 

1 s K\efyu cKAtTreTo) ^ fpx *]DK (Est. 9" and * 73 19 seem to indi- 
cate that forms of ]io were read here; cf., however, DtK *]DK (I 8 ), 
where the verb was undoubtedly regarded as in the 3d person. 
The absence of a translation for SD makes it entirely uncertain 
what the Septuagintist read in his text). 1 s ircrara e|iy (collec- 
tive) ; dvo/xovs D1K (this is a change for theological reasons rather 
than an inner-Greek corruption from dv0/>o>7rous [cf. I 17 , avOp<i>- 
TTOVS DIK]. It is unnecessary to suppose that the Septuagintist 
had either D'yeh D1K [G. A. Smith] or [Gratz] D'KBn in his text). 
I 4 ovofuiTa DE? (collective) ; KM additional. I 5 Sw/xara nUJ (cf. * 
129"). I 6 aTTo nnD; KOL TOVS p) ^rowras lK?p3 X 1 ? ItfKl (the 
Greek and Hebrew differ in regard to the verb-form to be used 
with the negative) ; dn-cxo/Aecous TOV Kvpiov int^"n (the participial 
form is again retained; for the sake of clearness the suffix is 
translated by its logical antecedent, cf. Jer. 8 3 ). I 7 vAa/?er0 
DH (cf. Zech. 2 17 ); Ova-uiv avrov rUT (cf. 2 14 , avr^s). I 8 Kal lora* 
..... KOL e/cSt/cTjo-w s mp)l .... mm (the Septuagintist does not 
seem to have understood the Hebrew tense consecution) ; 
BhaSD (collective). I 9 TrpoVvXa jnso (cf. I 12 ) ; eov additional be- 
tween "JIK and its suffix (these words were thought to refer to 
the temple, and by this addition the reference is brought out 

1 Because the translator of Zephaniah seems to have known no law but caprice in his 
translations of the article, these have not been referred to ; for the same reason there are 
but few references to tenses. 

34 The Text of Zephaniah. 

more clearly). 1" e^p^e'vot ^DJ (cf. II Sam. 24 12 ) ; cv rrj 
run (cf. 3 16 ). I 12 Xv X vov rnu (cf. I 9 ). I 13 ev avrats additional. 
I 14 Initial on and Kat additional (interpretative). I 18 Kat o-TrovS^v 
nSn3J }K (^r^rai, Jer. 15 8 ). 2 2 opyrjv *]K pin (cf. Is. 13 13 and 
Jer. 4 8 , passages in which 0v/xos translate f]X pin . In 3 8 Dyr seems 
to have been omitted in translation because the Septuagintist's 
supply of synonyms was exhausted). 2 3 rairtwoi ^y (= Si, 3 1S ) ; 
OTTWS^^SIK. 2 4 Kat additional; eKpt^T/o-erat niBnr (the construc- 
tion is changed to avoid the resumptive suffix). 2 5 dXXo<vX(Dv 
(this is the usual translation except in the Hexateuch). 2 5 
v/xas sprnatfn (the object in the Septuagint is not Canaan, 
but the Philistines); CK KaroiKtas 3tfr pKD (cf. 3 6 ). 2 8 7rpo/3aY(ov 
pv (collective). 2 7 rots KaraXoiTrois rviKtf (concrete for abstract, 
cf. 3 13 ); eTrco-KCTTTat ips 1 (=KStK>7(ra>, I 8 - 9 ; the change of tense 
is interpretative). 2 8 ovetSto-^ovs r\3in (cf. 1 1S ) ; e/xeyaXwovro 
iVun (= e/xeyaA.w0*7<rav, 2 10 ) ; opia fiou oSuj (the reading of 
the Septuagint is intrinsically improbable, for the phrase 
my border in the sense of territory occurs nowhere else 
with Jehovah as speaker ; cf. I Chr. 4 10 ). 2 9 Kat (2) ad- 
ditional ; Kvptos Ttov 6\W/A<ov ni&OY mrr (= TravroKpaTw/o, 2 10 ) ; 
KaToAowroi nniy and in s . 2 10 toy omitted (this omission was inten- 
tional to strengthen the idea, cf. Jer. 48 aM2 ). 2 11 c7ri</>avvjo-cTat 
N1U (cf. Joel 2 31 , Hab. I 7 ) ; TWV eOv&v additional (this addition cor- 
responds to the change from gods to kings in the Peshitta). 2 14 
KCU' (4) and avnjs (4) additional; OrjpLa TI/S y^s 'U irrn (the 
Septuagint has the phrase as it occurs in Gen. I 24 , cf . * 79 2 ) ; 
Orjpia <j><j)vir)<Tei 111^ Sp (?) ; Kat ^a/xatXcovres Kat e^tvot 13p DJl HNp DJi 

(collectives). 2 15 cXTrt's nan (cf. EC. 9 4 ) ; vo/4 0>;ptW n^nS pin. 
3 2 Kat additional. 3 3 ev Dips (cf. 3 1M7 ). 3 3 w? (2) additional. 
3* Kat additional. 3 6 e^ep^/xwo-o) ^nninn (cf. 2 7 ). 3 6b and 2 13 are 
good illustrations of free and literal translation. 3 7 Kat additional ; 
^o\oOpvO'rfT ni:r (the re. was added under the influence of the 
preceding verbs) ; TrdVra oo-a SD . 3 8 VTro/xavov /xov ^ 13H ; dvaa-- 
/x.ov ''Dip ; ets (rvvaycoya? e0i/cov rov cts Se^acr^at ^SacrtXet? rov tK^eat 
^p 1 ? D'U ^DX 1 ? (^ao-tXets is a contraction for jSooiXe&t?) ; 
minus S D^T (cf. 2 2 ). 3 9 yXwoxrav HSt^; Travra? D^D (cf. I 7 , 2 8 ) ; 
vyov DDiy (change of figure). 3 10 CK Treparwv TTOTO/AWV 'At0tO7rtas 
1^13 nn: 1 ? 13^D (cf. 2 15 ). 3 13 Kat additional, mxtf is taken as the 

Interdependence of the Versions. 35 

subject of the preceding, not of the following, verb ; K<U OVK lorrai 
6 eK<o/3a>v avrovs Tina JW (the part, in this phrase never has an 
object, cf. Is. 17 a , etc.; cf. also 2 6 , 3", I 7 , 2"). 3 14 Bvyarep 'leppv- 
Stm?' (perhaps the change is due to the following w 
cf. Gen. 36 81 B.) ; 0X775 rr}? KapSuxs <rou n 1 ? Sm. 3 16 Xc- 
<re K x/3os l\^P^ v " ov ^P'** 3 (interpretative expansion ; 
it is unnecessary with Stekhoven to suppose that JH3 was read) ; 
h /xeVw o-ov }31p3 (cf. 3 3 ' 17 ). 3 16 v T<? Kcupo> OVD (cf. I 13 ). 3 17 
ev o-ot }:np:j (cf. 3 3 , 3 16 ); object pronoun additional (2). 3 19 Xe'y 
Kvptos additional. 3 19 oi/o/xacrrovs Dfc? (3 20 idem). 
additional ; eW>7riov V 



I. The Peshitta is of Post-Christian origin, and in New 
Testament times the Septuagint was already so well established 
that it was quoted as authoritative. The wide popularity that 
the Septuagint enjoyed would tend to cause many of its phrases, 
expressions and interpretations to pass into current use, and 
some of these may have been unconsciously adopted by the 
makers of the new version. As they were not entirely familiar 
with Hebrew, it is natural that they should consult the existing 
version when in doubt. The two translations continued to exist 
side by side as ecclesiastically recognized versions, and correction 
of the one by the other is therefore not impossible, especially 
since many Greek ecclesiastics were resident in Syria for a longer 
or shorter time. That the Peshitta and Syro-Hexapla may 
have influenced each other mutually is shown by the case of Bar- 
Hebraeus, who is known to have used them both. The probability 
of interdependence, either initial or subsequent, thus established 
is so strong that the Septuagint and the Peshitta cannot be 
regarded as independent witnesses when they agree together 
against the Hebrew. 1 In Zephaniah the influence of the Septua- 

1 This principle is of course invalid when the two versions follow a common tradition 
that can be located in the Aramaic Targum. The Targum of Zephaniah is, however, so 
paraphrastic that it gives little aid to Textual Criticism. That which it offers can here be 
conveniently collected:! 6 DD^D is interpreted as idols. I 9 "1J1 J^IH is explained as 
those who walk in the laws of the Philistines. 2* The imperatives are rendered by 
words having the root idea of assemble. 2 14 The Targum adds K3iy to ^Ip . 3 1 PHIID 
nJOID . 3 DD^ is represented by ^fQ 3" The obscure clause of this 1 verse is inter- 
preted by the captivity of my people ivhich was taken captive. 

36 The Text of Zephaniah. 

gint on the Peshitta (or vice versa, cf. 3 10 ) can be discerned in 
several places. 

I 7 0^0 evAa/?eto-0 (Syro-Hexapla, a^-*?). I 11 
Hwoi dpyv/oio) (Syro-Hexapla ^Nn4Sn). I 12 ^0019040^, ^ft . K ara- 
<j>povovvra<i CTTI TO, ^vXay/Aara avTtov (Syro-Hexapla, U'4J). I 13 v oous 
additional eV avrats additional (?) . 1 14 P--^c U .00 i-j^o VLKp ^ Ka i 
<rK\rjpa (reraKTat) Awa-n/ (Syro-Hexapla, P-^^ paa^io] ]. *-oo 
r^s o-a/oKas avrw. 2 1 1?> P? l^> ojJflj^o a-*lsA.l__ <r 
TO 0vos TO aTratScvTOV (Syro-Hexapla, 1|-^1 cjj0]A.]o 
5 P? V^a^). 2 2 ^0001^ U,^ Trpo Tot) yeveV0ai v/xas (Syro-Hexapla, 
The Peshitta and the Septuagint both omit or in trans- 
lation. 2 s P-? 0,^1^0 Kpfa epyde<r6e. 2 6 Ur^ Kprfrrj. 2 7 l^a- 
additional T^S Oa\do-<nj<s additional. SA - X ^1 eTri^anyo-eTat (cf. the 
Syriac translation of &nij in Joel 2 11>S1 and Hab. I 7 ) ; t-^aJ eoAo- 
Opevo-t. 2 14 ^oiJJ UQ-^* Qypia. ^vyja-u. 3 1 l^^r- eTri^avrjs (Syro- 
Hexapla, 1^^-t- 1 ) ; l^Q-r 3 aTroXfXvTpupivri (Syro-Hexapla, l^-t-s). 
3 s t-1 (2) additional <is (2) additional (?). 3 7 <*^?<t> pdcrOe (?) ; 

. 3 8 


rvpiov. 3 9 r* ir-^^ VTTO Cvyov Iva. 3 10 The Septuagint and the 
Peshitta both omit s vi} ra nnj;. 3 18 ^.oov^ additional avTovs addi- 
tional. 3 17 ^^r* /catvicio-e NO-^J ^] <J, S l v ^eaa. 3 19 

The use that the Peshitta translator made of the Septuagint is 
on the whole a very intelligent one, although agreement in error 
can be found in the above list (l ia , 3 1 ). In I 14 he preferred to omit 
DSP rather than accept TeVaKTeu. In 2 3 he refused to accept a7roKpiW0c 
avTa, but he adopted the Septuagint interpretation of the first clauses 
of the verse ; dependence on the Septuagint was responsible for the 
omission of the second wpl. Perhaps the translation of p^na 
(2 14 ) by the colorless -^f> is due to the Septuagint Siopvy/Aao-iv. 
The Peshitta has hardly any demonstrable departures from the 
Massoretic tradition which it has not derived from the Septuagint. 
That the Peshitta has influenced the Septuagint in some of the 
instances cited, while not impossible, is still extremely doubtful. 
There is, however, to be found in many Greek manuscripts a very 
early translation of the obscure phrase in 3 10 (7rpoo-8e'o/xcu ev &e- 
o-TrapjuteVots /AOV) ; perhaps this belonged to the original Septua- 
gint, but was omitted under Syriac influence. In at least some 

Interdependence of the Versions. 37 

of the readings cited the Peshitta seems to have influenced the 
Syriac translation from the Septuagint (cf. 2 1 , 3 1 ). 

II. Jerome lamented the fact that in his day the world ' was 
divided between three opposing texts of the Septuagint.' 1 It was 
his purpose in his translation to get behind the Septuagint back 
to the "Hebrew verity"; and though he frequently reminds his 
reader that his work is not condemnatory of the ancients, 2 he is 
not slow to point out wherein and how they erred. As the pur- 
pose of Origen was similar to his own, he was naturally a great 
admirer of the Hexapla. His use of it can readily be illustrated 
by a few quotations : 

2 7 Quod autem legitur in ' Septuagintam ' a facie filiorum Juda, 
obelo praenotavimus, nee in Hebraeo enim, nee apud ullam fertur 
interpretatum. 3 9 Ubi nos interpretati sumus reddam populis 
labium electum, pro electo 4 Septuagintes ' dixerunt in generationem 
ejus, ut subaudiatur, terrae. Et hinc error exortus est, quod 
verbum Hebraicum BARURA, quod Aq. et Theo. electum, Sym. 
Mundum interpretatus est, 'Septuagintes' legerunt BADURA. 8 

3 18 Miror autem Aq. et ' Septuaginta ' in eo loco ubi 

diximus: congregabo quia ex te erant, pro erant interpretati 
voluisse vae, sive ot, quod semper Aq. non pro plangendo, sed pro 
vocando et inclamando ponit. 

There are only a few passages in the Vulgate which seem to 
indicate direct dependence on the Septuagint. Nomina and 
6v6fjuara (I 4 ), silete and evXafttLo-Oe (1 7 )> transeuntem and Tropcvo/xevov 
(2 2 ), et attenuabit and /cat eoA.o0/oaxrei (2 11 ), the additions of quasi 
and d>s (3 3 ), expecta and vTro'/xeii/ov (3 8 ), and the addition of the 
same suffix (3 9 ) may all be accidental. Columba (3 1 ) may not be 
due to the Septuagint ircpto-Tcpa, for in Jer. 25 38 Jerome makes the 
same mistake. Corvus (2 14 ) is, however, an intentional agree- 
ment with the Septuagint against the Hebrew of his day (Quod 

1 Totus orbis hac inter se trifaria varietate compugnat. Preface to Chronicles. 

2 Obsecro te lector ne laborem meum reprehensionem existimes antiquorum. 

9 In Zephaniah none of the minor Greek Versions are extant in manuscript, and only 
fragments contained in quotations such as these have been recovered. The longest of 
these fragments is one from Symmachus, preserved by Theodore of Mopsuestia : (3 10 ) 
TrtpaOev irorafjiuv ''AiBioirtas iKer&jovrd /we TKVO. T&V 8ie<TKOpTrur/>v vir' 1 tfwv tvty- 
K<a<ri Swpov tfjiot. They are of no textual value, for they have for the most part been 
preserved in citation only because they agree with the Massoretic Text against the Sep- 

38 The Text of Zephaniah. 

nos et * Septuaginta ' similiter transtulimus corvus in 

Hebraeo ponitur HAREB. The Vulgate and the Septuagint 
agree further in the peculiar addition of TiSK between <I JIK and 
its suffix (I 9 ) ; in the interpretation of *pDJ (2 1 ), 101:1 (2 s ) and 
nSjUJ (3 1 ) ; and in the subordination of the independent clause 
of 3 ao . All these agreements can hardly be accidental, 'especially 
since it is known that Jerome was thoroughly familiar with the 



I. Vowels and accents were introduced into the Hebrew 
text not earlier than the sixth century A. D. The so-called 
Sopherim in the first Christian centuries fixed the form of writing 
as regards the matres lectionis. There is good reason to believe 
that there was neither word or sentence division in the earliest 
manuscripts. It is therefore possible to consider the consonantal 
text entirely apart from the form which tradition has given it by 
the word and verse division that now obtains, and from the inter- 
pretation that the vocalization gives it; for these are as it were 
superimposed upon the original text. The translators in some 
cases adopted a possible reading or interpretation which disagrees 
with the Massoretic tradition. J 

(a). 1 s Congregans (2) t|OK (* efoKj cf. I 2 ). I 5 Melchom 
D^Sp (*o3Sp is read in the Massoretic text only in I Kgs. 11 B ' 8 
and II Kgs. 23 13 ; it has been proposed to read it also in I Kgs. 
II 7 , II Sam. 12 30 , I Chr. 20 2 , Am. I 15 and Jer. 49 1 ' 3 . Jehovah is 
often called a king [cf. 3 1B ; Harper, Amos and Hosea, p. 141; 
and Brown, Driver and Briggs' Hebrew Lexicon, Art. ^D, 3], 
The name Milchom occurs or is proposed as a reading only in 

1 In this and the following lists the readings that must be rejected as evidently wrong 
have been indicated by a star. Some that have only the negative merit of being not 
impossible have been left undistinguished, and those that commend themselves some- 
what more strongly have been marked with a dagger. 

Departures from Massoretic Tradition. 39 

passages in which Ammon is mentioned in the immediate context. 
In this passage there is no reference to Ammon, and therefore it 
is necessary to accept the Massoretic punctuation and to look for 
the exact meaning of the words in the use of different preposi- 
tions with the same verb). l u conticuit nmj (* n? ?"^ T ). I 14 The 
Vulgate takes in as a predicate adjective with Dr. 2 & perditorum 
DWD fD'in?, cf. I Kgs. 7 12 Pro CHORETIM, quod dicitur, 
perditorum, nomen Cretae Insulae (' Septuagintes ') putaverunt 
Jerome). 2 14 attenuabo mix (*H^, cf. 2"). 3 8 in futurum 
~W^ (t Ubi nos transtulimus, in die resurrectionis meae in 
futurum, et omnes interpretati sunt, in testimonium, Hebraeus, 
qui me in Scripturis instituit, asserebat LAED in praesenti loco 
magis ek eri, id est, in futurum debere intelligi Jerome; cf. Is. 
30 8 , Am. I 11 , Is. 9 8 and Gen. 49" in the Vulgate). 3 18 Nugas 
;u (*nugas .... a nobis ita ut in Hebraeo erat positum, ut nosse 
possimus linguam Hebraeicam omnium linguarum esse matricem. 
This has been characterized as an interesting bit of crude com- 
parative linguistics. Amara = in (I 14 ) would have served 
Jerome's purpose much better. The Septuagint offers a parallel 
in x<w>s=\J, (Mic. I 4 ). 3 18 ut non ultra habeas Sy fiKtyn (* L ?j; n&u?n, 
cf. Lev. 19 7 ). 3 20 in tempore quo congregabo ^3p AP3 (f AP?). 
See further under (c) 2 2 , 3 1 ' 1 , 3 20 . 

(5) ! 6 >onn\sn D5Sn (cf. Vulgate, supra). I 11 ojo^nnij (*naru). 
2 14 V^H* 3in (^l 1 ?)- 3 n Kinn DVD is connected with what precedes 
and not with what follows. See further under (c) I 11 , 2 a , 3 1 , 3 8 , 3 19 . 

(c) I 1 TOV TOV xovo-i 'Bho p (the Septuagintist has interpreted 
'J^O 1 3 patronymically as the following mov seems to show). I 11 
rv)v KaraKCKo/x/oieVr/v ^roDH (* i^nDQn) ; kfjioiuOrj npnj (there are two 
similar roots of which one means * to be like', the other, to destroy). 
I 12 <v\ay/wrra (cf. Peshitta) avrwv Drnni?( * DnnD, cf. Ex. 12 42 ). 
I 14 reraKrat D (*Dt?) ; ^13 J was read as an adjective with the fol- 
lowing DV and not as a noun. I 17 KCU expect }3i?i (* ^3^,1). 2 a 
TTopevd/xevov (cf. Peshitta and Vulgate) "O;? (j- 13 j;). 2* aTro/c/oiVeo-- 
^c avra nij^ (* ^nwjj). 2 14 Stopvy/xacm/ j^n (this meaning of the 
word is to be found nowhere else in the Septuagint). 3 1 aTroXeXv- 
Tpa>/u,eV>7 (cf. Vulgate and Peshitta) nS&u: (there are two roots 
SKJ, of which one means * to redeem ; the other, to pollute) ff 
(cf. Vulgate). mrn (this form may be a * noun subject 

40 The Text of Zephaniah. 

of nyntf, 3 2 ; or a participle from nr, to oppress). 3 3 'Apa/?ias 
anj; (* 3lj, cf. Jer. 25"; in Hab. I 8 the Septuagint bas the same 
reading, and in Jer. 5 6 ecus TWV ofciW represents rn3ij? = rr:i iy). 3 
(cf. Peshitta) ItfS (i#7; cf. Yulgate, supra). 3 12 nnKtf 
is taken as the subject of the preceding verb by the Septu- 
agint. 3 16 epct Kvpos "iraK?. (* lOK" 1 was read and KV/HOS was inter- 
pretatively added, cf. 3 19 ). 3 19 ev o-ot eve/cei/ o-ou ipyo ho fitf (* }fiK 
W?S) ; T^V KTTTTLe^vr}v (cf. Peshitta) j^n (cf. Mic. 4 8 ' 7 , Gen. 
32 33 ; there are two roots yStf of which the y is represented in Arabic 
by Dad and Ta respectively; the one may mean, to oppress-, the 
other, to limp. The former is found in the Old Testament only 
in the noun ySv, rib ; but the Septuagint suggests very plausibly 
that the word here and in Mic. 4 6 - 7 be taken from the root that 
has the meaning of to oppress. Barth [ Wurzeluntersuchungen . . . , 


pp. 39, 40] suggests the Arabic J^ as the cognate of the word 
here used). 3 19 h rw Kai/ow orav eto-Sc'lo/xat '^p '"W? (f AP?, of. 

(d) These readings, so far as they are not at once condemned 
by internal evidence, are suggestive for the interpretation of the 
text. Whether they are wrong or right, they shed additional 
light on the translations and translators. They also indicate the 
gradual crystallization of the tradition that grew up around the 
text, for departures from it decrease in the versions in the order 
of their rise. 

II. The readings of the versions so far considered either 
agree with the current consonantal text, or else data are lacking 
to show that the text of which they are severally the translation 
varied from it. It is still necessary to consider the equivalents 
in regard to which there is positive evidence of disagreement. 
Many variants are by the context or by parallel passages shown 
to be due to the intentional or unintentional faithlessness of the 
translators to their copies, or to the defective character of the 
exemplars which they used. Those that are not thus condemned 
will represent each version's actual contribution to the textual 
criticism of Zephaniah. 

(1) Most of the variants are reducible to the addition, omis- 
sion, transposition, or change of single letter?. 

Variants from Consonantal Text. 41 

(a) I 14 tribulabitur rm (I 1 *, I 17 and especially Am. 3 11 show 
that *[n]iY was read; the rendering by the future is interpretative, 
cf. Am. 3", idem). 2 14 robur e/wa my (*n ; Ty, cf. Pr. 21", Jer. 
5 1 68 ; this change may be due to the punctuation of the preceding 
nnK which Jerome adopted). 3 1 provocatrix HfcOlD ( Quod signi- 

ficantius Hebraeice dicitur *MARA, id est, TrapairiKpaivov<ra 
Jerome). 3 10 filii r3 ('J3; this may, however, be an inner-Latin 
corruption from t /jfo'a). See further under (c) 2 14 . 

(b) ).f\.. -ppjn (* n^pSn). I 10 )r>$ D>:n (OTV, this word is 
always so rendered in the Peshitta, when it occurs in connection 
with lytf). See further under (c) 2 a - 2 , 3 1 , 3 7 - 7 , 3 17 . 

(c) I 8 OLKOV "J3 (confusion between JVD and "ft is frequent; O'KOS 
= ', Jer. 16 16 , Ez. 2', I Chr. 2 10 ; vloi=no, Gen. 45", Ex. 16", 
Jos. 17 17 , 18 16 , Hos. I 7 ). I 9 e/M<av(us jiSin (it has been suggested 
that ^n or rhl was read ; there may, however, be a corruption in 
Greek here. Perhaps a participial form of e/u./ftuvo), i. e., e/A/?avras, 
stood in the original Septuagint; cf. I Sam. 5 6 , Greek). I 10 cbro- 
KO/TOWTCDV D'jnn (* D'jnn ; cf. II Chr. 33 14 ). I 14 o-KA^pd m* (the Sep- 
tuagintist has translated a derivative of "ny, to be hard). 2 a irpb 
TOV yivsvOai v/xa? (cf. Peshitta, Pr^^zK 1 ? Qiwa) pn trh D"D3 (per- 
haps the Septuagintist readiprnn K^ D*iD3 and made use of the color- 
less yeve'o-00,1 to translate the verb because he misread po) ; av0os 
^D (*p) ; the omission of or from the Septuagint (cf . Peshitta) 
may be due to a reading DOT'iDy. 2 5 irdpoiKoi "U (Pro GOI, id est, 
gente, l Septuagintes ' legerunt * GAR, hoc est, advenam 
Jerome). 2 14 KopaKs (cf. Vulgate) :nn (t^ij?) ; TO Avao-riy/xa avr^s 
my (*nvy, cf. Jer. 15 8 , Hos. II 9 , * 73 20 ; my, to rowse, is ren- 
dered by firavLa-rrjfjiL in Job 17 8 ). 3 1 (TTi^av^ (cf. Peshitta and Syro- 
Hexapla) HKIID (* ^, cf. 2 11 ). 3 8 4ir W ^vow D' (f D ' w ). 
3 7 ^ 6<t>0a\fji>v a.vTrjs (cf . Peshitta) miyD (f n^ryo) ; eToi/xaov (cf. 

Peshitta) opOpicrov tyOaprai ira(ra ^ eTri^vXXts avrwv IDOtfn |3K 
omVSy ^73 in^ni^n (cf . the Vulgate rendering of the two verbs ; 
* orviSSy SD in^nt^n opt^n pn) 3 9 ets ycveav avr^s mna (* nTna)^ 
3 12 evXaprjOTjo-ovTat, ion(* 1DH, cf. I 7 and Ne. 8"; 0. = ! is due to 
this reading). 3 17 firdu vw (* rnsr; KCUVW? (cf. Peshitta) nn' 
(f E^Hn"). 3 18 TOVS o-wrerpt/x/xevovs ^JOD ( D"pp ; with erov in the 
Greek text, ^3n, cf . Ex. 5 16 ) ; 3 18 ovac rn (f 'in, cf. 2 5 ) ; rts 
nfe/D (f KET" % n). 

42 The Text of Zephaniah. 

(2). There are a few readings which suggest a somewhat 
greater difference of text than those just considered do. 

1* ovojjLara iKtf (oty, perhaps this is due to the DK? in the context, 
or to either Hos. 2 19 or Zech. 13 a ). 3 18 d>s ev ^e/oa (cf. Peshitta) 

Qvja-ovrai DHtf 3 (* B'0^?rj ; for the sake of variety this was rendered 
by a passive, since DT\DJy occurs in the immediate context). 3 20 

(3). The words in the versions for which there is no equiv- 
alent in the Hebrew are, as has already been indicated, evident 
expansions of an interpretative character. Where the Hebrew 
text is fuller than that reflected by a version, explanation is not 
so easy. The only word not represented in the Vulgate is mD 
(2 6 ). Except where it is dependent on the Septuagint (I 14 , 2 a , 2% 
3 10 ?), the Peshitta text is as full as the Hebrew text with but a 
single insignificant exception (2 12 , a suffix and a demonstrative pro- 
noun omitted). There are only a few places in which the Septu- 
agint has no equivalent for words to be found in the Hebrew text. 
Decision as to whether these words convict the Septuagintist of 
omission or illustrate the "growth of the Massoretic Text" must 
from the nature of the case be largely subjective. From the 
time of Luther scholars have remarked a tendency on the part of 
the translators of the Septuagint to omit what they did not under- 
stand. The translator of Zephaniah must be charged with omis- 
sion on this score. 

1* Dp D'lDDn . Chemarim occurs in only two other places in the 
Old Testament. In one of these the Septuagint transliterates it 
(II Kgs. 23 5 ), and in the other its translation is the result of a 
transparently inappropriate etymology (maji KM Ka0o>s 
Kpavav avrdv, Hos. 10 6 ). It is thus entirely probable that this (and 
the following) word was omitted because it was not understood. 

I 6 . The only argument that can be advanced in regard to 
D'lnntfon (2) is stylistic. It seems to make the construction 
rather awkward. Cod. Q omits D'jntfjn (2) ; this may be the 
Hesychian reading, for Cod. Q is an Egyptian manuscript. This 
disagreement in the Septuagint makes it difficult to determine 
which word, if either, is additional in the current Hebrew. 

2 s . A desire to make the construction uniform may account 
for the omission of lute and the suffix of infltfn. The mis- 

Variants from Consonantal Text. 43 

reading of nup, by which the Peshitta was led astray, seems to 
have caused the omission of the third uppu. This verse is a good 
illustration of how the Septuagint influenced the Peshitta. 

3 10 . '13 nanny. The meaning of these words is not clear, and 
it is therefore more probable that they were omitted by the 
Septuagintist than that they were interpolated into the Hebrew 
subsequent to the time of translation. If 7rpoo-8e'o/xai lv Sieo-Tra/o/Ac- 
vots fiov, as has already been suggested, was in the original Septu- 
agint, it was later omitted under Syriac influence. 
may be a corruption for Tr/aoo-ei^o/xai (cf. Ju. 13 8 ). ev 
fjiov seems to represent 'Jfisp? (cf. II Chr. 18 18 ), which agrees 
closely with the reading that the Yulgate seems to suggest (filii 
dispersorum meorum "^ \^). 

(4). In the three passages that remain to be considered the 
Hebrew is difficult, and help from the versions would be very 

I 3 , et ruinae impiorum erunt V*4-*'V^ |A1 ]A^oculo Ka l do-^cv- 
ricrova-Lv ol cure/Jets D'yenn n m^EODrn. The versions all agree 
as to D"ych (cf. Peshitta, Num. 16 ao ), but each one gives it a dif- 
ferent grammatical government. They also agree in regard to 
the root htiD (cf. Septuagint, Ez. 21 20 ), though not in regard to the 
form of it here to be read. The n, which is difficult, is not 
represented in the Septuagint or Yulgate, and the Peshitta seems 
to have read it as the first person imperfect of nnx. Jerome 
wrote among other things in regard to these words, pro quo Sym. 
interpretatus est, et scandala cum impiis, ut subaudiatur, con- 
gregabuntur, sive deficient' Quinta autem ed., et infirmitas cum 
impiis deficiet. It would seem from this quotation that Jerome 
knew of the r\K in the text, and that the Vulgate translation 
is supposed to do justice to it. Though it is quite certain that 
this troublesome word is not represented in the Septuagint, it is 
impossible to determine what the Greek does represent. Perhaps 
the first word was read as a perfect with vau conversive; on this 
supposition the Hebrew has sometimes been corrected. The 
witness of the versions is contradictory and entirely inconclu- 

2 6 . Eteritfuniculus maris requies pastorum, et caulae pecorum 

44 The Text of Zephaniah. 

7rot/mW KCU /jiavSpa Trpo/Jarw |l3f niTUl D'JH HID nu O'H 
Whether n^n or nrrn was read by the translators, it is impossible 
to determine. The Vulgate has omitted mo and read nnj for 
nu (requies is the constant translation of nnj). The Peshitta has 
interpreted the verse freely in accordance with its reading of mD 
(Crete) adopted from the Septuagint. DTI Snn is not represented 
in the Septuagint; the order of nu and mj is reversed; mj is 
read as a proper name; irot/mW translates D^n (cf. 2 14 iroipvui=; 
D'Yiy; at vofuu TO>V woi/xva>v=D'jnn 111 JO, Am. l a , seems to indicate 
that 7rot/AV6o>v must be corrected to Troi/xeVwv) . Perhaps the addi- 
tion of TT}S 0a\d<ro"r)s (2 7 ) is compensatory for the omission of 
DTI S^n (cf. Peshitta). The difficulty of the translators seems to 
have focussed in mj, which is a hapax legomenon. This word 
is by many regarded as a gloss on nu ; to others both it and DTI ^n 
seem superfluous. The impossibility of correcting the Hebrew 
by the versions is patent, but Slin nrrn is inexplicable (the noun 
is always masculine except in this verse, cf. 2 7 ). The wide 
divergence of the versions from the current Hebrew and from 
each other becomes clear when the various texts are placed side 
by side in translation : 

(a) Revised Version And the sea-coast shall be pastures, with 
cottages (caves; others, wells) for shepherds and folds for flocks. 

(b) Vulgate And the sea-coast shall be a place of rest (cf. 
Verg. A. Ill, 393) for shepherds, and a fold for sheep: 

(c) Peshitta And the sea-coast shall be a dwelling place, and 
Crete a pasture for flocks of sheep : 

(d) Septuagint And Crete shall be a pasture for flocks, and a 
fold for cattle. 

2 8 . siccitas spinarum et acervi salis r^l , *oi-^J & S n tiZ|? 
.^quMn \Sr Aa/Aao-Kos cKAeAei/AjueV?; a>? ^t/xwna dAds fl^D mDDl SlIH pK^DD. 
Jerome read p.Eto, and acerviis dependent on Olivia (Siccitas, quod 
Hebraeice MAMASAC; . . . MEM si mutetur et DALETH 
accipiatur, easdem litteras habet quas et Damascus ; . . . 0i//,<uviav, 
id est, acervum). The Peshitta seems to have read m^D (Job 
30 4 ), mallow. Snn (Syriac, IV*; cf. Prov. 24 31 , Lee, li-^~) was 
in contrast with mSn, which grows wild, interpreted as cultivated 
grasses. n"OD was read as a passive form of n"O and translated 
as always by r^. A parallelism was produced by giving the 

Conclusion. 45 

remaining word a corresponding meaning. The relative and suf- 
fixes, as well as the conjunction and copula, of the next verse are 
interpretative additions. The Syriac is thus to be rendered: 
because their crop has been destroyed, and their wild grass has 
perished. The origin of Aa/xao-Kos is explained by Jerome. e/cAc- 
Xufjt,fj,tvri shows that Sin was read for Sin. 0i/Awvwx. shows that 
n"OD was derived from jvo in a sense preserved in the Aramaic 
('"O, UH>) and the Assyrian (karu). oXos must, as has already been 
indicated, be read for dAon/os. While witnessing to the orginality 
of the current Hebrew, the versions give absolutely no help in its 



Everything in the versions that seemed to have a bearing on 
the criticism of the text has now been presented with as much 
fulness as it seemed to warrant. The nature of the material con- 
sidered makes differences of opinion in regard to its proper dis- 
tribution inevitable, but the necessity for some such scheme of 
classification as has been adopted will hardly be denied. The 
departures of the Vulgate from the Massoretic tradition which 
have been noted have no special merit, and of the readings in 
which it bears positive witness to a difference between its 
"Urtext" and the present Hebrew not one is worthy of con- 
sideration. In every case its witness to the text on which it is 
based (cf. I 14 - 2"), or the witness of that text itself (cf. S'^is 
unreliable. The Peshitta, when it is independent of the Septua- 
gint, disagrees with the Massoretic tradition very infrequently, 
and the few variants that it offers are no more worthy of accept- 
ance than are those of the Vulgate. So far as it can be con- 
trolled, the testimony of these two versions is in favor of the 
accurate transmission of the Hebrew from the time of their 
origin. This conclusion would perhaps need some revision, if 
the numerous non sequiturs due to the process of translation 
could be eliminated. To possess the manuscript or manuscripts 
used by the translators would therefore be of considerable advan- 
tage to Textual Criticism. 

46 The Text of Zephaniah. 

If the recovery of the sources of the Vulgate and the Peshitta 
is a thing to be desired, the possession of the source of the Septu- 
agint is positively a sine qua non for the full understanding of the 
history of the Hebrew text of Zephaniah, for this translation is 
but a sorry equivalent for its original. It was not made by one 
who had a "genius for translation", for his general inaccuracy 
seems to have been even greater than his lack of knowledge, 
unless indeed he attempted to cover his ignorance by manipulat- 
ing his text. Many of his translations call vividly to mind the 
hit or miss achievements of a school-boy whose pensum stands 
between him and the play-ground. Luther accused the Septua- 
gintists, as a body, of "disdaining to speak the letters, words 
and style". To show the justice of this criticism as far as 
Zephaniah is concerned, one need only to remove the numerous 
faulty or wrong translations and interpretations from the Greek 
text; for hardly a verse will then remain intact. A comparison 
of the possible with the impossible variants in the consonantal 
text that it definitely supports shows that the attitude which 
must be maintained toward the Septuagint of Zephaniah is one of 
general distrust. It rarely agrees with the Massoretic text, 
where that text is difficult ; but the alternates which it suggests 
are generally even less acceptable. 1 It cannot be appealed to as 
an infallible authority on hapax legomena, nor can the Hebrew 
lexicon be enriched by the meanings of rare words that it sup- 
ports. Since the testimony of the Septuagint as to its source is so 
unreliable, its value for Textual Criticism is much less than it 
might be in view of the comparative nearness of its " Urtext " to 
the autograph. It is especially unfortunate in this case that the 
Septuagint does not speak with a more certain voice either in 
condemnation or confirmation because of the difficulties which 
the Hebrew presents. 2 The only general conclusion warranted 
by the facts is that the Septuagint offers no conclusive evidence 
that the " Lagardian archetype" was not the text on which it 

for B^IT (3 17 ) has gained wide acceptance, and yet against this possibly 
correct reading three positively wrong readings of T or "| must be balanced in this book. 

2 The difficulties in the Hebrew and in the interpretation of Zephaniah are briefly pre- 
sented in Appendix I. 

Difficulties in the Hebrew Text. 47 

also was based. As far as the possibility of showing the con- 
trary by external evidence goes, the present Hebrew text may 
well be that of the autograph of Zephaniah, 1 for the few parallels 
in thought and diction with other parts of the Old Testament to 
be found in the book are of no critical value (I 6 Jer. 8 2 ; 1" 
Jer. 48 11 ; I 13 Am. 5 n ; I 18 Ez. 7 lfl ; 2 8 Is. 16', Jer. 48 ae - 48 ; 2 14 - 1& 
Is. 13, 21 - 22 , 34 11 , 47 8 - 10 ; 3 4 Ez. 22 28 ; 3 10 Is. I 18 ), and the versions 
offer not a single reading which absolutely demands acceptance. 



The words and phrases included in this list have occasioned a 
great deal of discussion. It may be safely affirmed that in regard 
to them nothing is certain. 

l a 'JDK ^DK. The infinitive absolute is from a different root than 
the finite verb. (sjDKK, Wellhausen; *]?', Nowack; cf. * 104", 
Mi. 4 B ). 

I 9 rK. The word stands between two nouns (r\K 'n^Eon, Oort). 

I 5 D\j?3tfan D'mntfon. The juxtaposition of these two participles 
is awkward. (Some would omit the former, while others prefer 
to delete the latter). Ehrlich (Mikrd Ki Pheschuto, III, p. 456) 
suggests that the use of different formulas of swearing is indi- 
cated by "a j?3Bfo and "h yzwi ; the former referring to the s n form, 
the latter to the ^K form. 

I 9 |r2D hy jSin. The Targum seems to connect the words with 
the custom of the Philistine worshippers of Dagon, I Sam. I 6 ; cf. 
Trumbull, The Threshold Covenant, 2d ed., p. 117. Ehrlich (p. 
457) translates: die in denVorzimmern herumscharwenzelen. He 
thinks that sycophants are referred to, and that they are com- 
pared to dogs leaping up and down at the threshold of their 

I 14 nnn. This word must be read as a participle (IHDD, Well- 

1 The protests of Conjectural Criticism and Higher Criticism do not properly fall 
within the limits of the present inquiry, but a few remarks which seem not entirely 
uncalled for have been added in Appendix II. 

48 The Text of Zephaniah. 

I 14 "11:2:1 rm in rnrr or Sp. The grammatical relation of these 
words to each other and to what precedes is obscure. (The con- 
jecture of Gratz is rather heroic, 113:0 niir nirr ^p). 

1 17 DnS. The exact meaning of this word is unknown (cf. Job 

1 18 nSmj. nSrVj is the ordinary form. 

2 1 iKhpi itfuhpnn. The meaning of the words is unknown. 

pDJ is also uncertain (the Aramaic *|DD means turn pale). 

2 a D1D3 with an infinitive occurs only here (in Is. 17 14 and 28* 
it is used with a noun), and the pleonastic use of S with this con- 
junction is found nowhere else in the Old Testament. 

2 8 The word ^n seems to be feminine in this verse ; in the next 
verse it is masculine, rnj is found only in this verse; the usual 
form is m*O. niD is a hapax legomenon of doubtful meaning 
(Ehrlich, nyi?). 

2 7 It is not clear to whom the suffix of orr 1 ?;? refers (D^n hy, 

2 9 p$DD and niDD are hapax legomena, and the meanings usually 
given to the words are conjectural. 

2 11 The tense of nil is difficult and its meaning is obscure. 

2 14 nip nrw o ^DU nn p^na intf* Sp? Ehrlich suggests that the 
3 of 3in is due to dittography, and he translates the first five 
words: es pfeift lustig zum Fenster hinein, zum Loch an den 

3 1 In HK1D the is hard to explain. 

3 8 1D1J is by many regarded as a hapax legomenon (cf . Septua- 
gint), others take the word as a denominative from DiJ (cf. Nu. 
24 8 , Ez. 23 24 ). 

3 4 ni1J3 is a hapax legomenon as to form. 

3* 11V3 is a hapax legomenon. 

3 7 "U1 Sj seems to hang in the air. (It has been proposed to 
read n^jJD with the Septuagint, to change rn:r to ino% and to 
take hy s nip) in the sense of command, Lagarde.) 

3 10 'injj is a hapax legomenon. 'ysna ? 

3 17 "3 a^irv. A direct object for the verb seems necessary 
OTO, cf. * 21 7 ). 

3 18 There are two roots to which uu may be referred; of these 
one means to be grieved, the other, to be removed. The two 

Criticism of the Text. 49 

translations offered by the Revised Version illustrate the extreme 
obscurity of this verse. 

3 19 nx-niyy is unusual (Gratz suggests that n^D be added, cf. 

The grammatical governments of unwi is not clear (Noldeke 
proposed to delete the final D of D'nDfr and to take Dntfa as its 



I. No one can say what may or may not happen to a text 
transmitted in manuscript, and therefore not even the wildest 
conjecture can be dismissed as impossible; but it is equally true, 
even though the contrary seems to be implied in the confident 
assertions of some, that the fact that Zephaniah may have expressed 
a thought in a certain form or written a sentence in a certain 
way does not actually prove that he did so write or express it. 
The relative plausibility of the readings which it has been pro- 
posed to substitute for those in the current Hebrew can be more 
or less accurately gauged. In Appendix I the conjectures that 
have something positive to recommend them have already been 
noted. A free reconstruction of the text obtained by raising 
poetical measure 1 or the demands of a fantastic theory 3 into a 
canon of Textual Criticism has hardly more validity than have 
the results of an entirely arbitrary change, transposition and 
recombination of letters. 3 The changes which show only what 

1 Much study has been devoted to Hebrew poetry in the last two decades. Miiller 
(Die Propheten in Hirer ursprunglichen Form; Strophenbau und Responsion), 
Konig (Stilistik, Rhetorik, Poetik) and Sievers (Studien zur Hebraischen Metrik) 
have contributed largely to the recent popularity of this subject. The latest attempt to 
recast Zephaniah in poetical form was contributed by Fagnani to the Harper Memo- 
rial Volumes (1908). 

a Cheyne (Critica JBiblia, in loc.) has changed 2* to read: D'BO HIIT 1 ? 
DTJp I 1 ? Uni. He has the following note in support of one of his changes: 
is required as a parallel to lS though represented only by } in IBhpl. 

' Bachmann(^wr Textkritik des Propheten Zephanja, S.K ; 1894) has emended 
to read : m^D .... HD33 vh 'UH- 

50 The Text of Zephaniah. 

the critic thinks Zephaniah ought to have said can with safety be 
dismissed from serious consideration. 1 

II. This free Conjectual Criticism of the text gives much 
support to and gains much help from the Higher Criticism, which 
dissects an ancient document according to subjective standards 
of style and thought-cogency. The integrity of Zephaniah 
has often been denied. The following summary condensed from 
the article Zephaniah by J. A. Selbie in Hastings' Dictionary of 
the J3ible needs very little comment. 2 Keunen was inclined to 
regard 3 1 *- 90 as post-exilic on account of differences both in tone 
and situation from the rest of the prophecy. Stade denied to 
Zephaniah 2 1 - 8 - 11 and the whole of chapter 3. Wellhausen (com- 
pare Nowack) suspected 2 2 - 3 , rejected 2 8 " 11 and treated chapter 3 
as a later supplement added in two stages (1-7 and 8-20). Budde 
(followed by Cornill, Einleitung, 3d edition) admitted 2 1 ' 1 , 
3 1 " 6 ' 7 ' 8 ' 6 ' 11 " 18 as in harmony with Zephaniah's situation; he rejected 
2 4 " 16 mainly because Israel appears as the victim, not as the per- 
petrator of wrong ; he excluded 3 9 ' 10 as breaking the connection 
between 3" and 3" ; he declared 3 14 " ao to be a later lyrical epilogue. 
Schwally allowed to Zephaniah chapter 1, 2 13 ' 16 and perhaps 2 1 ' 4 , 
holding 2 6 " 1 * to be exilic and chapter 3 post-exilic, though 3'" T 
may be Zephaniah's. G. A. Smith denied to Zephaniah 2 8 " 11 , 
3 9 ' 10 and 3 14 " 30 . Driver remarked that 2 11 seemed to be somewhat 
out of place and that 3 14 " 20 is somewhat doubtful, though the 
' question remains whether it is sufficiently clear that the imagina- 
tive picture was beyond the power of Zephaniah to construct.' 
Davidson defended the genuineness of chapter 2 as a whole, but 
considered it quite possible that it had been expanded in various 
places; he allowed that 3 10 should possibly be omitted, but other- 
wise 3 1 " 13 appeared to him to be genuine, although they might 
suggest that the passage was later than chapter 1 ; in 3 14 " 20 he 
recognized quite a different situation from the rest of the book. 
Konig would apparently accept the whole of the book except the 
title which refers the prophecy to the days of Josiah. 

This paragraph is an unintended, though on that account no 
less positive, refutation of the method by which such conflicting 

1 D^TJ? for Dmy and v^OY for niBP (2' 4) are of this kind. 

2 The article Zephaniah in the Encyclopaedia Bibllca contains a similar summary by 

Criticism of the Text. 51 

results are achieved. One can hardly repress the thought that a 
great deal of these " assured results" is due to the endeavor of 
each latest critic to justify his rediscussion of the subject by 
presenting something different from that which his predecessors 
have said. It would seem from this paragraph that the book in 
its present form is but a sorry piece of patchwork ; and yet the 
writer of the article Zephaniah in Smithes Dictionary of the Bible 
expressed the opinion that "the chief characteristics of this book 
are the unity and harmony of the composition, the grace, energy 
and dignity of its style, and the rapid and effective alternations 
of threats and promises." The critics themselves being wit- 
nesses, there is not a single verse which Zephaniah could not 
have written, and therefore one who is not anxious to father any- 
thing new can defend the integrity of the book by choosing his 
"authorities" with discrimination. The writer is free to con- 
fess that he is interested in the whole text, which may be Zepha- 
niah's Zephaniah, rather than in that part of it which in the 
opinion of each critic a Zephaniah, who was on the plane of 
religious evolution which he thinks his age had attained, who 
possessed the mentality with which he is pleased to endow him, and 
who wrote as he himself would have written under similar cir- 
cumstances, could or ought to have produced. The arguments 
and counter-arguments advanced for and against the genuineness 
of the many verses discussed are all singularly pointless and are 
invalid to overthrow the presumption established in favor of the 
integrity of the book by the mere fact that some one gave it its 
present form ; for to that man's mind the book was a unit and 
the ease with which critics brush aside the arguments of critics 
demonstrates that an unbiased Higher Criticism can not show 
that the man in question was not the Zephaniah to whom the 
book has so long been attributed. Arguments based on the style 
of a writer known only through his works are notably precarious, 
even though he has left extensive literary remains. The psycho- 
logical law of the Association of Ideas utterly condemns all 
argumentation based on thought development alone, for it shows 
that no combination or contrast of ideas even abrupt change from 
threat to promise is impossible. Zephaniah has left at most 
fifty-three verses ; it is surely absurd to build up one's conception 

52 The Text of Zephaniah. 

of the man out of the first eighteen that are assumed to be his, 
and to use the conception of his style and capacities thus gained 
as a standard to determine which of the remaining verses he 
could and which he could not have written. Judged by present 
standards, strong arguments can be advanced to show that 3'' 5b 
originally stood between the two halves of I 13 : 

(a). In the present text it is difficult to determine where the 
arraignment of Nineveh ends and that of Jerusalem begins. The 
Peshitta has actually referred 3 1 to Nineveh, and the present 
chapter division of the Septuagint shows that 2 15 was referred to 
Jerusalem by its author. 

(b). The nexus between the second and third clauses of 3 6 does 
not seem to be very close, but 3 Bc in that it would emphasize the 
absolute hopelessness of Nineveh's condition would be an admir- 
able conclusion to 2 15 . 

(c). 3 1 continues in the style of I 11 and 3 2 " 5b contain the full 
charge on which the punishment threatened in I 12b is based. 

The ipsi dixerunt of the critics have no greater objective 
validity than those for this transposition have. A detailed dis- 
cussion of all the points involved in this seemingly endless dis- 
cussion would lead far into the theory of Israel's religious 
development, whose exigencies seem to demand such excisions 
(2 3 - 11 , 3 8 " 11 ) as are not based on purely subjective considerations, 
and therefore the reader who seeks for arguments of this kind to 
support his belief in the integrity of the book must be left to 
find them in the works of such champions as each verse or verse- 
group has found. 1 

i The present tendency to find wholesale interpolations in the Prophets has been dis- 
cussed by Vos (The Eighth Century Prophets, Presbyterian and Reformed Review, 

x VITA. 

The writer was born in Meedhuizen, Province of Groningen, Holland, Janu- 
ary 25, 1883. He received his primary education in the Public Schools of 
Chicago, 111., and was graduated from Hope College, Holland, Mich., with 
the degree of Bachelor of Arts in 1903. He attended the sessions of the 
Princeton Theological Seminary during the years 1903-1907, receiving the 
Degree of Bachelor of Divinity in 1907. He was the Newberry Scholar of the 
Board of Education of the Presbyterian Church from 1904 to 1907. From 
1907 to 1909 he was a student in Columbia University, one year as a Fellow 
in Semitic languages. While in Columbia University he attended Old Testa- 
ment lectures in Union Theological Seminary, New York.